fix(fleet): dedicated retirement capability, identity proof, external fencing

Addresses the three blocking findings in review 657 on PR #982 and nothing
else.

B1 — apply was authorized by gitea.read
---------------------------------------
Plan and apply shared the observational permission class, so any profile that
could look could also destroy; the mutation landing in the local control-plane
registry rather than in Gitea makes it no less a mutation.

Apply now requires its own capability, gitea.worker_registry.retire, checked at
entry and re-resolved immediately before the registry mutation. Plan stays on
gitea.read. No profile holds the new permission by default, so author,
reviewer, merger, and read-only profiles fail closed on the permission itself
rather than on the role check alone; the controller/reconciler role restriction
remains as defence in depth. Granting it is a deliberate operator edit to
profiles.json, and removing it revokes apply completely. No new Gitea write
permission is introduced and no author permission is broadened.

B2 — complete legacy rows could be retired without identity proof
-----------------------------------------------------------------
"Incomplete identity" previously meant only that pre-existing columns were
null, which a legacy-pid row satisfies trivially. Retirement now requires an
affirmative two-part proof: launcher-minted inst- attribution, plus fencing
evidence (host_id, boot_id, process_start_time) that turns a bare pid into a
statement about one process incarnation.

New mcp_process_fencing supplies those probes; every one returns None rather
than guessing, and None always preserves. Rows written before these columns
existed, and rows on legacy instance identities, are preserved permanently —
they are retired only after re-registering under a trusted identity. That
legacy rows would otherwise remain outstanding is explicitly not treated as
grounds for a weaker proof. A live pid stays an absolute block even across a
boot boundary, and pid reuse is reported distinctly from a live worker.

B3 — lease and OS liveness sat outside the registry transaction
---------------------------------------------------------------
BEGIN IMMEDIATE locks the worker registry only, and the per-target loop runs
after revalidation, so a lease acquired or a pid revived in between would go
unnoticed — the registry-column guard cannot catch it because no registry
column changed.

Two mechanisms now close that window, both applied per target immediately
before its own write: an external_fence_fn version token over active leases,
captured inside the transaction before the authoritative read and re-compared
before every guarded UPDATE (movement aborts the whole transaction; an
unreadable lease store raises rather than comparing equal), and a liveness_fn
re-probe that must affirmatively re-establish that this exact process is gone,
comparing process_start_time so a reused pid is refused. Omitting the re-probe
retires nothing rather than proceeding unfenced. The guarded UPDATE also
asserts the fencing triple is unchanged, and all three columns participate in
the CAS token.

Preserved from the accepted work: registry_fingerprint still excludes
observation time and is order- and numeric-typing stable, plan still mutates
nothing, drift still retires zero, retired rows stay historical, and ordinary
author operations remain ungated by fleet state.

Tests: tests/test_issue_980_stale_worker_retirement.py 87 passed, 46 subtests
(was 40 passed, 23 subtests), covering all three blockers' required
regressions. Adjacent suites (#980, #978, #975, #948, task-capability role
invariants) 242 passed, 156 subtests. Full suite from the branch worktree
28 failed, 6253 passed, 6 skipped, 1152 subtests; master baseline at
108cbfa173 from branches/baseline-980-108cbfa 28 failed, 6166 passed, 6
skipped, 1106 subtests. The failing sets are identical under comm, so zero
regressions and zero masked pre-existing failures; the +87 passing delta is
this branch's tests.

No live worker-registry row was retired — every test uses a throwaway SQLite
database. BAA, issue #981, PR #906, issue #650, review 624, unrelated branches
and worktrees, profiles, configuration, credentials, sessions, and running
processes were not touched.

Refs #980
Co-Authored-By: Claude Opus 4.8 (1M context) <[email protected]>

canonical_repository_root.py

@@ -37,6 +37,48 @@ CANONICAL_ROOT_ENV = "GITEA_CANONICAL_REPOSITORY_ROOT"
 # Candidate git remote names probed when deriving repository identity.
 _IDENTITY_REMOTE_CANDIDATES = ("prgs", "origin", "dadeschools", "mdcps")
 
+# Repository-authority modes (#973 B10). Exactly two values are supported.
+# ``mode`` selects how repository authority is established, so an unrecognised
+# value must never be normalised onto one of these: aliasing a trusted mode is
+# precisely the defect. Omitting the argument keeps the documented safe default,
+# ``validation``.
+MODE_VALIDATION = "validation"
+MODE_DERIVATION = "derivation"
+SUPPORTED_MODES: tuple[str, ...] = (MODE_VALIDATION, MODE_DERIVATION)
+
+# Reason code emitted when an unsupported mode is refused. Mirrors the existing
+# ``webui.sanctioned_restart.DENY_UNKNOWN_MODE`` convention so callers and tests
+# can assert the refusal cause rather than string-matching prose.
+DENY_UNKNOWN_MODE = "unknown_mode"
+
+
+def unsupported_mode_reason(mode: object) -> str | None:
+    """Precise rejection reason for *mode*, or None when *mode* is supported.
+
+    Only the two documented string values are accepted, compared exactly — no
+    stripping, no case folding — so misspellings and whitespace variants are
+    refused rather than coerced. An empty string, ``None``, and any non-string
+    are all *explicitly supplied* unsupported values and are refused on the same
+    footing; none of them is normalised to a supported mode. Omitting the
+    argument entirely never reaches here with an unsupported value because the
+    parameter default is ``"validation"``.
+    """
+    if isinstance(mode, str) and mode in SUPPORTED_MODES:
+        return None
+    supported = ", ".join(repr(m) for m in SUPPORTED_MODES)
+    if not isinstance(mode, str):
+        return (
+            f"unsupported repository-authority mode {mode!r} of type "
+            f"{type(mode).__name__}: only {supported} are supported; the mode "
+            "is refused before any repository assessment and no repository "
+            "identity was resolved through it (fail closed)"
+        )
+    return (
+        f"unsupported repository-authority mode {mode!r}: only {supported} are "
+        "supported; the mode is refused before any repository assessment and no "
+        "repository identity was resolved through it (fail closed)"
+    )
+
 
 def configured_canonical_root(
     profile: Mapping | None,
@@ -132,6 +174,7 @@ def assess_canonical_repository_root(
     process_project_root: str,
     remote: str | None = None,
     require_binding: bool = False,
+    mode: str = "validation",
 ) -> dict:
     """Validate the canonical repository root binding, failing closed on forgery.
 
@@ -140,15 +183,41 @@ def assess_canonical_repository_root(
     ``configured`` (whether a cross-repo binding was declared),
     ``resolved_slug`` and ``source``.
 
-    Without a configured binding the single-repo default is preserved: the
-    canonical root is derived from *process_project_root* and never blocks
-    (unless *require_binding* explicitly demands one).
+    *mode* accepts exactly the two values in :data:`SUPPORTED_MODES`:
+    - ``"validation"`` (default): an independently trusted expected repository
+      slug is known (or required). Candidate root's observed identity must match.
+      Unprovable or missing expected identity fails closed when *require_binding* is True.
+    - ``"derivation"``: caller is deriving the canonical repository identity.
+      No expected slug exists yet by design. Derivation succeeds if the configured
+      root exists, is a git repository, and carries a resolvable git remote identity.
 
-    With a configured binding the path must exist, be a git repository, and —
-    when *expected_slug* is known — carry a matching repository identity. A
-    mismatched or (when *require_binding*) unprovable identity is a forged or
-    conflicting binding and fails closed.
+    Every other explicitly supplied value — unknown strings, misspellings, the
+    empty string, ``None``, and non-strings — is refused with ``proven`` False,
+    ``block`` True, and ``reason_code`` :data:`DENY_UNKNOWN_MODE` (#973 B10).
+    Omitting *mode* entirely keeps the documented ``"validation"`` default.
     """
+    # #973 B10: refuse an unsupported repository-authority mode as the very first
+    # act, before any candidate-root existence check, path or symlink resolution,
+    # git top-level discovery, remote-URL or repository-identity discovery, and
+    # before any expected-versus-observed comparison or validation/derivation
+    # behaviour. An unsupported mode previously fell into the catch-all ``else``
+    # on the configured-root path (silently receiving validation semantics) and
+    # skipped the identity comparison entirely on the single-repository default
+    # path (strictly weaker than validation), so matching identities could return
+    # ``proven`` True. No repository identity may be resolved through a mode the
+    # contract does not define.
+    mode_reason = unsupported_mode_reason(mode)
+    if mode_reason is not None:
+        return _assessment(
+            proven=False,
+            reasons=[mode_reason],
+            configured=bool((configured_value or "").strip()),
+            canonical_repo_root=None,
+            resolved_slug=None,
+            source=source,
+            reason_code=DENY_UNKNOWN_MODE,
+        )
+
     process_root = os.path.realpath(process_project_root)
     declared = (configured_value or "").strip()
 
@@ -168,12 +237,31 @@ def assess_canonical_repository_root(
             )
         # Single-repo default: canonical root follows the install checkout.
         derived = resolve_repo_toplevel(process_root) or process_root
+        resolved_slug = repository_identity_slug(derived, remote=remote)
+        reasons: list[str] = []
+        # #973 B10: the allowlist above guarantees *mode* is one of the two
+        # supported values here, so an unsupported value can no longer skip this
+        # identity comparison and end up strictly weaker than validation.
+        if mode == MODE_VALIDATION and expected_slug:
+            expected = expected_slug.strip()
+            if resolved_slug and resolved_slug.lower() != expected.lower():
+                reasons.append(
+                    f"canonical repository root identity mismatch: '{derived}' resolves "
+                    f"to repository '{resolved_slug}' but expected repository identity "
+                    f"is '{expected}' (forged or conflicting binding, fail closed)"
+                )
+            elif not resolved_slug and require_binding:
+                reasons.append(
+                    f"canonical repository root '{derived}' has no resolvable git "
+                    f"remote identity to confirm authorization for '{expected}' "
+                    "(fail closed)"
+                )
         return _assessment(
-            proven=True,
-            reasons=[],
+            proven=not reasons,
+            reasons=reasons,
             configured=False,
             canonical_repo_root=derived,
-            resolved_slug=None,
+            resolved_slug=resolved_slug,
             source=None,
         )
 
@@ -207,19 +295,36 @@ def assess_canonical_repository_root(
 
     resolved_slug = repository_identity_slug(toplevel, remote=remote)
     reasons: list[str] = []
-    expected = (expected_slug or "").strip() or None
-    if expected:
-        if resolved_slug and resolved_slug.lower() != expected.lower():
+
+    if mode == MODE_DERIVATION:
+        if not resolved_slug:
             reasons.append(
-                f"canonical repository root identity mismatch: '{toplevel}' resolves "
-                f"to repository '{resolved_slug}' but the session is authorized for "
-                f"'{expected}' (forged or conflicting binding, fail closed)"
+                f"configured canonical repository root '{toplevel}' has no resolvable "
+                "git remote identity (fail closed)"
             )
-        elif not resolved_slug and require_binding:
+    else:
+        # #973 B10: MODE_VALIDATION only. This arm is no longer a catch-all — the
+        # allowlist above admits no third value, so an unsupported mode can no
+        # longer silently receive validation semantics here.
+        expected = (expected_slug or "").strip() or None
+        if expected:
+            if resolved_slug and resolved_slug.lower() != expected.lower():
+                reasons.append(
+                    f"canonical repository root identity mismatch: '{toplevel}' resolves "
+                    f"to repository '{resolved_slug}' but the session is authorized for "
+                    f"'{expected}' (forged or conflicting binding, fail closed)"
+                )
+            elif not resolved_slug and require_binding:
+                reasons.append(
+                    f"canonical repository root '{toplevel}' has no resolvable git "
+                    f"remote identity to confirm authorization for '{expected}' "
+                    "(fail closed)"
+                )
+        elif require_binding:
             reasons.append(
-                f"canonical repository root '{toplevel}' has no resolvable git "
-                f"remote identity to confirm authorization for '{expected}' "
-                "(fail closed)"
+                f"canonical repository root '{toplevel}' has configured value "
+                f"'{configured_value}' but authoritative expected repository identity "
+                "is unprovable or missing (fail closed)"
             )
 
     return _assessment(
@@ -252,10 +357,19 @@ def _assessment(
     proven: bool,
     reasons: list[str],
     configured: bool,
-    canonical_repo_root: str,
+    canonical_repo_root: str | None,
     resolved_slug: str | None,
     source: str | None,
+    reason_code: str | None = None,
 ) -> dict:
+    """Build the assessment payload.
+
+    ``canonical_repo_root`` is None only when the assessment refused to resolve
+    one at all — today exactly the unsupported-mode refusal (#973 B10), which
+    must not derive a trusted repository identity through an undefined mode.
+    ``reason_code`` is a machine-checkable refusal cause; None for every
+    ordinary (non-coded) outcome.
+    """
     return {
         "proven": proven,
         "block": not proven,
@@ -264,4 +378,5 @@ def _assessment(
         "canonical_repo_root": canonical_repo_root,
         "resolved_slug": resolved_slug,
         "source": source,
+        "reason_code": reason_code,
     }

control_plane_db.py

@@ -27,12 +27,12 @@ import uuid
 from contextlib import contextmanager
 from dataclasses import dataclass
 from datetime import datetime, timedelta, timezone
-from typing import Any, Iterator, Mapping, Sequence
+from typing import Any, Iterator, Sequence
 
 import dependency_graph
 import gitea_audit
 
-SCHEMA_VERSION = 6
+SCHEMA_VERSION = 5
 
 # Assignable work kinds only — raw monitoring incidents are never work items.
 WORK_KINDS = frozenset({"issue", "pr"})
@@ -280,6 +280,22 @@ def _ts(dt: datetime | None = None) -> str:
     return value.astimezone(timezone.utc).replace(microsecond=0).isoformat().replace("+00:00", "Z")
 
 
+def _realpath_or_raw(value: str | None) -> str:
+    """Normalize a filesystem path for compare-and-swap equality (#970).
+
+    Symlinks and ``..`` segments must not make two spellings of the same path
+    look different, but an unresolvable path must still compare as itself
+    rather than collapsing to empty — an empty result means "no path given".
+    """
+    text = (value or "").strip()
+    if not text:
+        return ""
+    try:
+        return os.path.realpath(os.path.abspath(text))
+    except Exception:
+        return text
+
+
 def _parse_ts(value: str | None) -> datetime | None:
     if not value:
         return None
@@ -419,7 +435,6 @@ class ControlPlaneDB:
                 self._migrate_incident_links_null_scope(conn)
                 self._migrate_lease_lifecycle_columns(conn)
                 self._migrate_session_ownership_columns(conn)
-                self._migrate_session_lifecycle_columns(conn)
                 self._migrate_usage_events_table(conn)
                 conn.execute(
                     "INSERT OR REPLACE INTO schema_meta(key, value) VALUES (?, ?)",
@@ -813,7 +828,6 @@ class ControlPlaneDB:
         pid: int | None = None,
         status: str = "active",
         controller_instance_id: str | None = None,
-        owner_process_started_at: str | None = None,
     ) -> dict[str, Any]:
         """Register/refresh a session row.
 
@@ -823,22 +837,16 @@ class ControlPlaneDB:
         controller instance can. It is never overwritten with ``None``, so a
         heartbeat from a caller that does not supply one cannot erase
         ownership.
-
-        *owner_process_started_at* (#969) is the OS start time of the owner
-        process at registration time. When present it is retained across
-        heartbeats (never cleared by ``None``) so later PID-reuse checks do
-        not depend on a live ``ps`` probe of a long-dead process.
         """
         now = _ts()
         instance = (controller_instance_id or "").strip() or None
-        proc_start = (owner_process_started_at or "").strip() or None
         with self._tx() as conn:
             existing = conn.execute(
                 "SELECT session_id FROM sessions WHERE session_id = ?",
                 (session_id,),
             ).fetchone()
             if existing:
-                if instance is None and proc_start is None:
+                if instance is None:
                     conn.execute(
                         """
                         UPDATE sessions
@@ -848,21 +856,7 @@ class ControlPlaneDB:
                         """,
                         (role, profile, namespace, pid, now, status, session_id),
                     )
-                elif instance is None:
-                    conn.execute(
-                        """
-                        UPDATE sessions
-                        SET role = ?, profile = ?, namespace = ?, pid = ?,
-                            last_heartbeat_at = ?, status = ?,
-                            owner_process_started_at = COALESCE(?, owner_process_started_at)
-                        WHERE session_id = ?
-                        """,
-                        (
-                            role, profile, namespace, pid, now, status,
-                            proc_start, session_id,
-                        ),
-                    )
-                elif proc_start is None:
+                else:
                     conn.execute(
                         """
                         UPDATE sessions
@@ -876,33 +870,18 @@ class ControlPlaneDB:
                             instance, session_id,
                         ),
                     )
-                else:
-                    conn.execute(
-                        """
-                        UPDATE sessions
-                        SET role = ?, profile = ?, namespace = ?, pid = ?,
-                            last_heartbeat_at = ?, status = ?,
-                            controller_instance_id = ?,
-                            owner_process_started_at = COALESCE(?, owner_process_started_at)
-                        WHERE session_id = ?
-                        """,
-                        (
-                            role, profile, namespace, pid, now, status,
-                            instance, proc_start, session_id,
-                        ),
-                    )
             else:
                 conn.execute(
                     """
                     INSERT INTO sessions(
                         session_id, role, profile, namespace, pid,
                         started_at, last_heartbeat_at, status,
-                        controller_instance_id, owner_process_started_at
-                    ) VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?, ?)
+                        controller_instance_id
+                    ) VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?)
                     """,
                     (
                         session_id, role, profile, namespace, pid, now, now,
-                        status, instance, proc_start,
+                        status, instance,
                     ),
                 )
             row = conn.execute(
@@ -918,165 +897,6 @@ class ControlPlaneDB:
                 (_ts(), session_id),
             )
 
-    def retire_session(
-        self,
-        *,
-        session_id: str,
-        reason: str,
-        actor_session_id: str | None = None,
-        details: Mapping[str, Any] | None = None,
-        now: datetime | None = None,
-        terminal_status: str = "retired",
-    ) -> dict[str, Any]:
-        """Terminalize one session row if it is still non-terminal (#969).
-
-        CAS on non-terminal status: concurrent retirements of the same row
-        yield exactly one ``retired`` outcome and subsequent
-        ``already_terminal`` outcomes. Never deletes historical rows. Writes a
-        durable ``session_retired`` event for audit.
-        """
-        moment = _ts(now)
-        reason_s = (reason or "").strip() or "unspecified"
-        term = (terminal_status or "retired").strip().lower() or "retired"
-        terminal_set = {
-            "retired",
-            "ended",
-            "terminal",
-            "dead",
-            "stale",
-            "orphaned",
-        }
-        with self._tx() as conn:
-            row = conn.execute(
-                "SELECT * FROM sessions WHERE session_id = ?",
-                (session_id,),
-            ).fetchone()
-            if row is None:
-                return {
-                    "outcome": "missing",
-                    "reason": reason_s,
-                    "prior_status": None,
-                    "new_status": None,
-                    "details": {"session_id": session_id},
-                }
-            prior = dict(row)
-            prior_status = str(prior.get("status") or "").strip().lower()
-            if prior_status in terminal_set:
-                return {
-                    "outcome": "already_terminal",
-                    "reason": reason_s,
-                    "prior_status": prior_status,
-                    "new_status": prior_status,
-                    "details": {"session_id": session_id, "idempotent": True},
-                }
-
-            # Optional: refuse when an active lease still names this session.
-            active_lease = conn.execute(
-                """
-                SELECT lease_id, status, expires_at FROM leases
-                WHERE session_id = ? AND status = 'active'
-                LIMIT 1
-                """,
-                (session_id,),
-            ).fetchone()
-            if active_lease is not None:
-                return {
-                    "outcome": "blocked",
-                    "reason": "live_lease",
-                    "prior_status": prior_status,
-                    "new_status": prior_status,
-                    "details": {
-                        "session_id": session_id,
-                        "lease_id": active_lease["lease_id"],
-                        "blocker": "active_lease_row",
-                    },
-                }
-
-            cols = {
-                r[1] for r in conn.execute("PRAGMA table_info(sessions)").fetchall()
-            }
-            if "retired_at" in cols and "retire_reason" in cols:
-                conn.execute(
-                    """
-                    UPDATE sessions
-                    SET status = ?, retired_at = ?, retire_reason = ?
-                    WHERE session_id = ? AND status = ?
-                    """,
-                    (term, moment, reason_s, session_id, prior.get("status")),
-                )
-            else:
-                conn.execute(
-                    """
-                    UPDATE sessions
-                    SET status = ?
-                    WHERE session_id = ? AND status = ?
-                    """,
-                    (term, session_id, prior.get("status")),
-                )
-            changed = conn.execute(
-                "SELECT changes()"
-            ).fetchone()[0]
-            if not changed:
-                # Lost CAS race — re-read.
-                refreshed = conn.execute(
-                    "SELECT status FROM sessions WHERE session_id = ?",
-                    (session_id,),
-                ).fetchone()
-                cur = (
-                    str(refreshed["status"]).strip().lower()
-                    if refreshed is not None
-                    else None
-                )
-                return {
-                    "outcome": "already_terminal"
-                    if cur in terminal_set
-                    else "blocked",
-                    "reason": reason_s,
-                    "prior_status": prior_status,
-                    "new_status": cur,
-                    "details": {
-                        "session_id": session_id,
-                        "cas_lost": True,
-                    },
-                }
-
-            detail_payload: dict[str, Any] = {
-                "session_id": session_id,
-                "prior_status": prior_status,
-                "new_status": term,
-                "reason": reason_s,
-                "actor_session_id": actor_session_id,
-                "pid": prior.get("pid"),
-                "role": prior.get("role"),
-                "profile": prior.get("profile"),
-            }
-            if isinstance(details, Mapping):
-                for key, value in details.items():
-                    if key not in detail_payload:
-                        detail_payload[key] = value
-            message = (
-                f"session {session_id} retired reason={reason_s} "
-                f"prior_status={prior_status}"
-            )
-            # events.work_item_id is nullable; session retirement is not work-scoped.
-            conn.execute(
-                """
-                INSERT INTO events(work_item_id, event_type, message, created_at)
-                VALUES (NULL, 'session_retired', ?, ?)
-                """,
-                (message[:2000], moment),
-            )
-            # Also persist a structured JSON line in the message when short enough
-            # by appending a compact summary (full detail stays in return value /
-            # audit log; events.message is human-readable).
-            return {
-                "outcome": "retired",
-                "reason": reason_s,
-                "prior_status": prior_status,
-                "new_status": term,
-                "details": detail_payload,
-            }
-
     def list_sessions(
         self,
         *,
@@ -1835,29 +1655,6 @@ class ControlPlaneDB:
             if name not in cols:
                 conn.execute(f"ALTER TABLE sessions ADD COLUMN {name} {decl}")
 
-    _SESSION_LIFECYCLE_COLUMNS: tuple[tuple[str, str], ...] = (
-        ("owner_process_started_at", "TEXT"),
-        ("retired_at", "TEXT"),
-        ("retire_reason", "TEXT"),
-    )
-
-    def _migrate_session_lifecycle_columns(self, conn: sqlite3.Connection) -> None:
-        """Add session retirement / PID-reuse provenance columns (#969).
-
-        Additive and idempotent. Pre-existing rows migrate with NULL; retirement
-        fills ``retired_at`` / ``retire_reason``, and new upserts may record
-        ``owner_process_started_at`` for stronger identity checks.
-        """
-        cols = {
-            row[1]
-            for row in conn.execute("PRAGMA table_info(sessions)").fetchall()
-        }
-        if not cols:
-            return
-        for name, decl in self._SESSION_LIFECYCLE_COLUMNS:
-            if name not in cols:
-                conn.execute(f"ALTER TABLE sessions ADD COLUMN {name} {decl}")
-
     def list_active_claims(
         self,
         *,
@@ -2132,6 +1929,168 @@ class ControlPlaneDB:
                 ),
             )
 
+    def retire_lease_worktree_path(
+        self,
+        lease_id: str,
+        *,
+        expected_path: str | None = None,
+        expected_status: str | None = None,
+        expected_session_id: str | None = None,
+        expected_owner_pid: int | None = None,
+        reason: str = "missing_worktree_path_retired",
+    ) -> dict[str, Any]:
+        """Retire a missing worktree_path binding from a control-plane lease (#970).
+
+        Clears worktree_path on the lease row, updates provenance_json with
+        durable retirement audit proof, and writes a worktree_binding_retired
+        event.
+
+        The update is a compare-and-swap (#970 review 644 B1/B3): the caller
+        states the exact path it audited and, when known, the lease status,
+        owning session, and owner pid it classified against. Every stated value
+        must still match the stored row, and the ``UPDATE`` itself is keyed on
+        the stored ``worktree_path``, so a concurrent writer that moved or
+        replaced the binding between audit and apply loses the race instead of
+        having its value silently overwritten. A mismatch raises and mutates
+        nothing.
+
+        ``expected_path`` is mandatory: a retirement that does not name the path
+        it intends to clear cannot be safe against concurrent recreation.
+        """
+        now_s = _ts()
+        expected_norm = _realpath_or_raw(expected_path)
+        if not expected_norm:
+            raise ControlPlaneError(
+                f"cannot retire lease {lease_id} worktree_path: expected_path is "
+                "required for compare-and-swap retirement (fail closed)"
+            )
+        with self._tx() as conn:
+            cols = self._lease_columns(conn)
+            row = conn.execute(
+                "SELECT * FROM leases WHERE lease_id = ?",
+                (lease_id,),
+            ).fetchone()
+            if not row:
+                raise ControlPlaneError(f"unknown lease_id {lease_id}")
+
+            record = dict(row)
+            current_wt = (record.get("worktree_path") or "").strip()
+
+            if not current_wt:
+                # Idempotent: the binding this caller audited is already gone.
+                return {
+                    "lease_id": lease_id,
+                    "retired": False,
+                    "already_retired": True,
+                    "prior_worktree_path": "",
+                    "expected_worktree_path": expected_path,
+                    "reason": reason,
+                    "compare_and_swap": {
+                        "matched": True,
+                        "outcome": "already_retired",
+                    },
+                }
+
+            if _realpath_or_raw(current_wt) != expected_norm:
+                raise ControlPlaneError(
+                    f"cannot retire lease {lease_id} worktree_path: expected "
+                    f"'{expected_path}' does not match current '{current_wt}' "
+                    "(fail closed)"
+                )
+
+            for field, expected_value in (
+                ("status", expected_status),
+                ("session_id", expected_session_id),
+            ):
+                if expected_value is None:
+                    continue
+                current_value = record.get(field)
+                if str(current_value or "").strip() != str(expected_value).strip():
+                    raise ControlPlaneError(
+                        f"cannot retire lease {lease_id} worktree_path: lease "
+                        f"{field} changed since audit (expected "
+                        f"'{expected_value}', found '{current_value}'); fail closed"
+                    )
+
+            if expected_owner_pid is not None:
+                current_pid = record.get("owner_pid")
+                if current_pid is not None and int(current_pid) != int(expected_owner_pid):
+                    raise ControlPlaneError(
+                        f"cannot retire lease {lease_id} worktree_path: lease "
+                        f"owner_pid changed since audit (expected "
+                        f"{expected_owner_pid}, found {current_pid}); fail closed"
+                    )
+
+            # Parse and update provenance_json
+            raw_prov = record.get("provenance_json") or "{}"
+            try:
+                prov = json.loads(raw_prov) if isinstance(raw_prov, str) else dict(raw_prov)
+            except Exception:
+                prov = {}
+            if not isinstance(prov, dict):
+                prov = {}
+
+            prior_path = current_wt
+            prov.update({
+                "worktree_path_retired": True,
+                "retired_worktree_path": prior_path,
+                "retired_at": now_s,
+                "retirement_reason": reason,
+                "retired_from_status": record.get("status"),
+                "retired_from_session_id": record.get("session_id"),
+                "worktree_path": "",
+            })
+            prov_json = json.dumps(prov)
+
+            if "worktree_path" in cols:
+                # CAS: keyed on the exact stored path this caller audited.
+                cur = conn.execute(
+                    "UPDATE leases SET worktree_path = '', provenance_json = ? "
+                    "WHERE lease_id = ? AND worktree_path = ?",
+                    (prov_json, lease_id, record.get("worktree_path")),
+                )
+                if cur.rowcount != 1:
+                    raise ControlPlaneError(
+                        f"cannot retire lease {lease_id} worktree_path: "
+                        "compare-and-swap matched no row (concurrent change); "
+                        "fail closed"
+                    )
+            else:
+                conn.execute(
+                    "UPDATE leases SET provenance_json = ? WHERE lease_id = ?",
+                    (prov_json, lease_id),
+                )
+
+            conn.execute(
+                """
+                INSERT INTO events(work_item_id, event_type, message, created_at)
+                VALUES (?, 'worktree_binding_retired', ?, ?)
+                """,
+                (
+                    record["work_item_id"],
+                    f"lease {lease_id} worktree_path '{prior_path}' retired: {reason}",
+                    now_s,
+                ),
+            )
+
+            return {
+                "lease_id": lease_id,
+                "retired": True,
+                "already_retired": False,
+                "prior_worktree_path": prior_path,
+                "expected_worktree_path": expected_path,
+                "retired_at": now_s,
+                "reason": reason,
+                "compare_and_swap": {
+                    "matched": True,
+                    "outcome": "retired",
+                    "expected_status": expected_status,
+                    "expected_session_id": expected_session_id,
+                    "expected_owner_pid": expected_owner_pid,
+                },
+            }
+
+
     def abandon_lease(
         self,
         *,
@@ -3270,3 +3229,123 @@ class ControlPlaneDB:
             "live_lease_id": None if live_lease_id is None else str(live_lease_id),
             "reconcile_action": "reconcile_required" if stale else "safe_to_resume",
         }
+
+    def retire_session_checkpoint_worktree_path(
+        self,
+        session_id: str,
+        *,
+        checkpoint_id: str | None = None,
+        expected_path: str | None = None,
+        expected_status: str | None = None,
+        reason: str = "missing_worktree_path_retired",
+    ) -> dict[str, Any]:
+        """Retire a missing worktree_path from session_checkpoints (#970).
+
+        Compare-and-swap, mirroring :meth:`retire_lease_worktree_path` (#970
+        review 644 B3). ``expected_path`` names the exact stored path the caller
+        audited; the guarded ``UPDATE`` is keyed on that stored value, so a
+        checkpoint whose path was moved, replaced, or concurrently rewritten
+        after the audit is refused without mutation rather than blindly cleared.
+
+        Exactly one checkpoint row is targeted: by ``checkpoint_id`` when given,
+        otherwise by ``session_id``, which must identify a single row.
+        """
+        now_s = _ts()
+        expected_norm = _realpath_or_raw(expected_path)
+        if not expected_norm:
+            raise ControlPlaneError(
+                "cannot retire session checkpoint worktree_path: expected_path "
+                "is required for compare-and-swap retirement (fail closed)"
+            )
+        if not checkpoint_id and not (session_id or "").strip():
+            raise ControlPlaneError(
+                "cannot retire session checkpoint worktree_path: checkpoint_id "
+                "or session_id is required (fail closed)"
+            )
+
+        with self._tx() as conn:
+            if checkpoint_id:
+                selector_sql = "SELECT * FROM session_checkpoints WHERE checkpoint_id = ?"
+                selector_params: tuple[Any, ...] = (checkpoint_id,)
+                selector_desc = f"checkpoint_id '{checkpoint_id}'"
+            else:
+                selector_sql = "SELECT * FROM session_checkpoints WHERE session_id = ?"
+                selector_params = (session_id,)
+                selector_desc = f"session_id '{session_id}'"
+
+            rows = [dict(r) for r in conn.execute(selector_sql, selector_params).fetchall()]
+            if not rows:
+                raise ControlPlaneError(
+                    f"cannot retire session checkpoint worktree_path: no "
+                    f"checkpoint matches {selector_desc} (fail closed)"
+                )
+            if len(rows) > 1:
+                raise ControlPlaneError(
+                    f"cannot retire session checkpoint worktree_path: "
+                    f"{selector_desc} matches {len(rows)} checkpoints; supply an "
+                    "exact checkpoint_id (fail closed)"
+                )
+
+            record = rows[0]
+            target_checkpoint_id = record.get("checkpoint_id")
+            current_wt = (record.get("worktree_path") or "").strip()
+
+            if not current_wt:
+                # Idempotent: the binding this caller audited is already gone.
+                return {
+                    "session_id": session_id,
+                    "checkpoint_id": target_checkpoint_id,
+                    "retired": False,
+                    "already_retired": True,
+                    "prior_worktree_path": "",
+                    "expected_worktree_path": expected_path,
+                    "reason": reason,
+                    "compare_and_swap": {
+                        "matched": True,
+                        "outcome": "already_retired",
+                    },
+                }
+
+            if _realpath_or_raw(current_wt) != expected_norm:
+                raise ControlPlaneError(
+                    f"cannot retire session checkpoint worktree_path for "
+                    f"{selector_desc}: expected '{expected_path}' does not match "
+                    f"current '{current_wt}' (fail closed)"
+                )
+
+            if expected_status is not None:
+                current_status = record.get("status")
+                if str(current_status or "").strip() != str(expected_status).strip():
+                    raise ControlPlaneError(
+                        f"cannot retire session checkpoint worktree_path for "
+                        f"{selector_desc}: status changed since audit (expected "
+                        f"'{expected_status}', found '{current_status}'); fail closed"
+                    )
+
+            cur = conn.execute(
+                "UPDATE session_checkpoints SET worktree_path = '', updated_at = ? "
+                "WHERE checkpoint_id = ? AND worktree_path = ?",
+                (now_s, target_checkpoint_id, record.get("worktree_path")),
+            )
+            if cur.rowcount != 1:
+                raise ControlPlaneError(
+                    f"cannot retire session checkpoint worktree_path for "
+                    f"{selector_desc}: compare-and-swap matched no row "
+                    "(concurrent change); fail closed"
+                )
+
+            return {
+                "session_id": session_id,
+                "checkpoint_id": target_checkpoint_id,
+                "retired": True,
+                "already_retired": False,
+                "prior_worktree_path": current_wt,
+                "expected_worktree_path": expected_path,
+                "retired_at": now_s,
+                "reason": reason,
+                "compare_and_swap": {
+                    "matched": True,
+                    "outcome": "retired",
+                    "expected_status": expected_status,
+                },
+            }

docs/instance-fleet-identity.md

@@ -0,0 +1,203 @@
+# Instance-level fleet identity and health snapshots
+
+Issue **#978**. Companion primitives: **#948** (worker ownership), **#975**
+(heartbeat lifecycle), **#951** (restart receipts).
+
+## Why this exists
+
+The multi-instance fleet gate (#963) needs a *native* answer to:
+
+* which application launches are live;
+* which of the five namespace workers belong to each launch;
+* whether two Codex (or Claude, Grok, …) launches are distinct;
+* whether any live collision is real rather than a shared client type.
+
+Process tables, PID proximity, configuration files, and direct SQLite access
+are **not** production evidence. The sanctioned surface is the read-only MCP
+tool `gitea_snapshot_instance_fleet` on **controller** and **reconciler**
+namespaces.
+
+## Identity hierarchy
+
+| Identity | Scope | Who generates it | Lifetime |
+| --- | --- | --- | --- |
+| `client_type` | Application family (`codex`, `claude_code`, `gemini`, `grok`, …) | Launcher sets `GITEA_MCP_CLIENT` | Stable for the product |
+| `client_instance_id` | One running application launch | **Trusted launcher**, once per launch, as `GITEA_MCP_CLIENT_INSTANCE` | Fresh launch → new ID; reconnect of same launch → same ID; full app restart → new ID |
+| `fleet_run_id` | Operator-approved enrollment / canary cohort | Operator / controller sets `GITEA_MCP_FLEET_RUN_ID` | Duration of the approved rollout |
+| `namespace` | `author` \| `reviewer` \| `merger` \| `controller` \| `reconciler` | Profile / MCP server binding | Process lifetime |
+| `worker_id` / `worker_identity` | One namespace worker process | Runtime registry at first registration | New on worker restart; not reused |
+| `session_id` | Client session ownership | Launcher `GITEA_MCP_CLIENT_SESSION` or runtime | Session lifetime |
+| `generation_id` | One daemon launch | Runtime at process boot | New on process restart |
+| `process_identity` / PID | OS process | Runtime | Process lifetime |
+
+### Rules
+
+1. Multiple active instances **may** share the same `client_type`.
+2. Every application launch receives a **distinct** `client_instance_id`.
+3. All five namespace workers of one launch report the **same**
+   `client_instance_id`.
+4. Each namespace worker has a **distinct** `worker_identity`, process
+   identity, generation, and PID.
+5. Instance identity is **never** inferred from PID proximity, timestamps, or
+   client type alone.
+6. Live reuse of one `client_instance_id` with conflicting workers fails closed.
+7. Sharing only a profile or `client_type` is **not** a duplicate.
+
+This deliberately **replaces** any permanent `exactly_one_per_profile` fleet
+model (#949 assumption) as the operating rule for multi-instance fleets.
+
+## How five workers join one instance
+
+1. The host starts one application instance (for example one Codex session).
+2. The **production application launcher**
+   (`mcp_application_launcher.build_application_mcp_servers` /
+   `gitea_config.multi_namespace_launcher_entries`) mints exactly one
+   `client_instance_id` via `mcp_fleet_snapshot.generate_client_instance_id`
+   and injects the same env into every namespace worker:
+
+   ```text
+   GITEA_MCP_CLIENT=<client_type>
+   GITEA_MCP_CLIENT_INSTANCE=<client_instance_id>
+   GITEA_MCP_INSTANCE_PROVENANCE=trusted_launcher
+   GITEA_MCP_CLIENT_SESSION=<session_id>          # optional but recommended
+   GITEA_MCP_FLEET_RUN_ID=<enrollment id>         # when on an approved canary
+   GITEA_CLIENT_MANAGED=1
+   GITEA_MCP_PROFILE=<role-profile>
+   ```
+
+3. The MCP client starts the five namespace processes (`gitea-author`,
+   `gitea-reviewer`, `gitea-merger`, `gitea-controller`, `gitea-reconciler`)
+   from that generated `mcpServers` map — each entry carries the **same**
+   instance ID.
+4. Each worker registers once into the worker registry with its own
+   `worker_identity`, `namespace`, `generation_id`, and PID, but the shared
+   `client_instance_id`. Workers never mint a trusted instance ID themselves.
+
+Only IDs matching the launcher format `inst-<client>-<timestamp>-<digest>`
+are trusted. Missing, `legacy-pid-*` / `pid-*` placeholders, and ordinary
+user-supplied strings fail closed as untrusted and cannot authorize
+multi-instance fleet mutation safety.
+
+Worker reconnect (same process, same registration) keeps the instance ID.
+Worker restart (new process) mints a new worker identity and generation but
+must still receive the same `GITEA_MCP_CLIENT_INSTANCE` from the parent
+application if it is the same launch. Full application restart mints a new
+`client_instance_id` (call the launcher without a prior ID).
+
+### Mutation gate (instance-aware)
+
+The capability / runtime diagnostic gate
+(`_check_mcp_runtimes_diagnostics`) is **instance-aware**:
+
+* Two legitimate application instances that share a profile (same
+  `GITEA_MCP_PROFILE`) are allowed when each has a distinct trusted
+  `client_instance_id`.
+* More than one live worker for the same
+  `(client_instance_id, profile/namespace)` fails closed as a duplicate
+  namespace worker.
+* Multiple processes sharing a profile **without** trusted instance
+  evidence still fail closed (indistinguishable from a duplicate).
+
+## Approved fleet manifest
+
+An operator (or controller enrollment step) obtains an approved manifest as a
+list of expected instances, for example:
+
+```json
+{
+  "instances": [
+    {
+      "client_type": "codex",
+      "client_instance_id": "inst-codex-20260730T120000Z-abc123def456",
+      "fleet_run_id": "canary-963-2026-07-30",
+      "namespaces": ["author", "reviewer", "merger", "controller", "reconciler"]
+    },
+    {
+      "client_type": "codex",
+      "client_instance_id": "inst-codex-20260730T120100Z-fed654cba321",
+      "fleet_run_id": "canary-963-2026-07-30",
+      "namespaces": ["author", "reviewer", "merger", "controller", "reconciler"]
+    }
+  ]
+}
+```
+
+Pass that list as `expected_manifest` to `gitea_snapshot_instance_fleet`.
+When a manifest is supplied:
+
+* instances on the manifest but not live → `missing_expected`;
+* live instances not on the manifest → `unmanifested`;
+* both are **active blockers** for fleet-gate safety.
+
+Without a manifest, the snapshot still enumerates the live fleet and classifies
+identity collisions; it does not invent enrollment policy.
+
+## Snapshot consistency
+
+Each snapshot includes:
+
+* `snapshot_at` — UTC timestamp;
+* `consistency_token` / `registry_revision` — content digest over worker
+  identity, instance id, status, heartbeat, fencing, and generation.
+
+Two successive snapshots can prove heartbeat continuity via
+`mcp_fleet_snapshot.compare_snapshot_heartbeats`.
+
+## Classification (active vs historical)
+
+**Active blockers** (make `live_fleet_safe=false`):
+
+* missing expected instance;
+* unmanifested extra instance;
+* duplicate namespace worker within one instance;
+* live `client_instance_id` collision;
+* reused worker / session / generation / process / PID / fencing identity;
+* orphaned or unowned workers;
+* unknown client;
+* foreign-repository workers;
+* old-revision workers;
+* stale workers still marked active;
+* legacy incomplete instance identity.
+
+**Historical dead rows** (`status` released/superseded) are reported as
+`historical_dead` findings with `active_blocker=false`. They never
+automatically make the live fleet unsafe.
+
+## Fail-closed behaviour and recovery
+
+| Condition | Mutation safety | Diagnostic reads | Recovery |
+| --- | --- | --- | --- |
+| Two instances, same `client_type`, distinct IDs | Safe (if otherwise healthy) | Available | None needed |
+| Live reuse of one `client_instance_id` | Unsafe | Available | Stop the colliding launch or re-issue a distinct ID |
+| Two workers, same namespace, one instance | Unsafe | Available | Stop the extra worker |
+| Legacy registration without trusted instance ID | Unsafe for fleet mutation | Available | Relaunch with launcher-issued `GITEA_MCP_CLIENT_INSTANCE` |
+| Historical dead row only | Does not block alone | Available | No action required for fleet safety |
+| Registry unavailable | Snapshot fails closed | N/A | Repair registry path / reconnect namespaces |
+
+Incomplete or untrusted instance identity **cannot** authorize unsafe
+mutation. Diagnostic reads remain available where `gitea.read` allows.
+
+## Permissions
+
+* **Allowed:** `controller`, `reconciler` with `gitea.read`.
+* **Denied:** author, reviewer, merger (even with `gitea.read` for other tools).
+* **No new mutation permissions** are granted to any role.
+
+## Related surfaces
+
+* `mcp_worker_identity` — registry, worker identity, heartbeats (#948, #975).
+* `mcp_fleet_snapshot` — pure snapshot + classification (#978).
+* `gitea_snapshot_instance_fleet` — sanctioned MCP tool (#978).
+* `gitea_get_runtime_context` — single-process view (not fleet-wide).
+* [stale-worker-retirement.md](stale-worker-retirement.md) — CAS-protected
+  retirement of conclusively stale registrations (#980). Note that the
+  `registry_revision` this snapshot returns is **time-seeded** and unsuitable
+  for compare-and-swap; retirement derives its own stable
+  `registry_fingerprint` from registry content alone.
+
+## Non-goals
+
+* Starting the #963 canary.
+* Purging historical registry rows.
+* Preserving “exactly one process per profile” as the permanent model.
+* Using shell / process-table / SQLite inspection as production fleet evidence.

docs/mcp-tool-inventory.md

@@ -51,6 +51,7 @@ that gates each call, not which tools exist.
 - `gitea_adopt_merger_pr_lease`
 - `gitea_adopt_workflow_lease`
 - `gitea_allocate_next_work`
+- `gitea_apply_stale_worker_retirement`
 - `gitea_assess_already_landed_reconciliation`
 - `gitea_assess_conflict_fix_classification`
 - `gitea_assess_conflict_fix_push`
@@ -65,6 +66,7 @@ that gates each call, not which tools exist.
 - `gitea_assess_work_issue_duplicate`
 - `gitea_assess_worktree_cleanup_integrity`
 - `gitea_audit_config`
+- `gitea_audit_missing_worktree_bindings`
 - `gitea_audit_runtime_recovery_contamination`
 - `gitea_audit_stable_branch_contamination`
 - `gitea_audit_worktree_cleanup`
@@ -123,19 +125,24 @@ that gates each call, not which tools exist.
 - `gitea_observability_link_issue`
 - `gitea_observability_list_projects`
 - `gitea_observability_reconcile_incident`
+- `gitea_plan_stale_worker_retirement`
 - `gitea_post_heartbeat`
 - `gitea_publish_unpublished_issue_branch`
 - `gitea_quarantine_contaminated_review`
+- `gitea_rebind_dirty_same_claimant_author_session`
 - `gitea_reclaim_expired_workflow_lease`
+- `gitea_reconcile_after_restart`
 - `gitea_reconcile_already_landed_pr`
 - `gitea_reconcile_issue_claims`
 - `gitea_reconcile_merged_cleanups`
+- `gitea_reconcile_missing_worktree_bindings`
 - `gitea_reconcile_superseded_by_merged_pr`
 - `gitea_record_daemon_process_kill_attempt`
 - `gitea_record_irrecoverable_decision_lock_provenance`
 - `gitea_record_pre_review_command`
 - `gitea_record_shell_spawn_outcome`
 - `gitea_record_stable_branch_push_attempt`
+- `gitea_recover_dirty_orphaned_issue_worktree`
 - `gitea_recover_incomplete_bootstrap_lock`
 - `gitea_release_merger_pr_lease`
 - `gitea_release_reviewer_pr_lease`
@@ -154,6 +161,7 @@ that gates each call, not which tools exist.
 - `gitea_sentry_reconcile_issue`
 - `gitea_sentry_watchdog`
 - `gitea_set_issue_labels`
+- `gitea_snapshot_instance_fleet`
 - `gitea_submit_pr_review`
 - `gitea_update_pr_branch_by_merge`
 - `gitea_validate_review_final_report`

docs/post-restart-reconcile.md

@@ -19,11 +19,7 @@ The assessor classifies:
 
 - **service_health** — process healthy / parity mutation-safe
 - **clients** — connected client descriptors (optional inventory)
-- **sessions** — active session rows with dead or reused owner pids are
-  unresolved until retired via `#969` (`session_lifecycle` /
-  `apply_session_cleanup=true` on `gitea_reconcile_after_restart`, or
-  `gitea_retire_stale_workflow_sessions`). Live owners, live leases, and live
-  client-managed sessions are never retired.
+- **sessions** — active session rows with dead owner pids are unresolved
 - **checkpoints** — soft-depends on #660; skipped with reason when schema absent
 - **leases** — live control-plane leases after restart
 - **capabilities** — master-parity / stale-runtime (#610)

docs/stale-worker-retirement.md

@@ -0,0 +1,282 @@
+# Stale worker retirement (#980)
+
+The #978 instance-fleet snapshot made registry accuracy observable but
+deliberately read-only: a registry full of rows whose owning processes are long
+gone stays full. This document describes the sanctioned way to retire those
+rows — a dry-run-first, compare-and-swap-protected workflow available only to
+controller and reconciler namespaces.
+
+Related: [instance-fleet-identity.md](instance-fleet-identity.md) (#978),
+[post-restart-reconcile.md](post-restart-reconcile.md) (#662).
+
+## Why a dedicated registry token
+
+`mcp_fleet_snapshot._consistency_token` seeds its digest with `snapshot_at`,
+formatted at second precision. Its output — surfaced as `registry_revision` and
+`consistency_token` on the snapshot — therefore changes on **every call**, even
+when no registry row changed. Any compare-and-swap gated on it can never pass:
+a dry-run/apply cycle spanning more than one second aborts unconditionally.
+
+That token remains useful as an observation stamp and is unchanged. #980 adds a
+separate, *stable* token instead:
+
+| Token | Module | Derived from | Stable across time? |
+| --- | --- | --- | --- |
+| `registry_revision` / `consistency_token` | `mcp_fleet_snapshot` | `snapshot_at` + a subset of row fields | **No** — moves every second |
+| `registry_fingerprint` | `mcp_fleet_retirement` | canonical retirement-relevant row content only | **Yes** |
+| `candidate_fingerprint` | `mcp_fleet_retirement` | canonical content of the selected candidate rows | **Yes** |
+
+`registry_fingerprint` guarantees:
+
+* identical canonical registry contents always produce the same token, whenever
+  they are observed;
+* row iteration order never affects the token (serialized rows are sorted);
+* any retirement-relevant change moves it — row creation or deletion, identity
+  change, heartbeat or TTL change, ownership change, registration-state change,
+  PID change, or repository-binding change.
+
+The exact field set is `mcp_fleet_retirement.FINGERPRINT_FIELDS`. Deliberately
+excluded: `token_fingerprint` (credential-adjacent, never a retirement input),
+the four `*_revision` columns (revision drift is an independent restart concern
+and is not part of the eligibility conjunction), and the `retired_*` bookkeeping
+columns this feature adds. Numeric values are canonicalised, so a TTL that
+round-trips through SQLite as `900.0` hashes identically to `900`.
+
+## Eligibility — the conjunction
+
+A registration is retired only when **every** one of these holds. Any missing
+or contradictory evidence preserves the row.
+
+| Requirement | Preserve reason code when it fails |
+| --- | --- |
+| `status` is `active` | `already_terminal_registration` |
+| Every field the conjunction reads is present (`REQUIRED_IDENTITY_FIELDS`) | `incomplete_registry_identity` |
+| `last_heartbeat_at` parses as a UTC stamp | `unparsable_heartbeat` |
+| Worker is not live | `worker_live` |
+| PID probe returns a definite answer | `pid_liveness_unknown` |
+| PID probe says the process is gone | `pid_alive` |
+| Heartbeat has expired under the canonical TTL | `heartbeat_not_expired` |
+| `ownership_state` is exactly `stale` | `ambiguous_ownership_state` |
+| Repository binding present and canonical | `repository_binding_ambiguous` |
+| No identity evidence shared with a live or unprobeable worker | `conflicting_identity_evidence` |
+| No other active row claims the same (instance, namespace) while one may be live | `client_instance_conflict` |
+| Not an active workflow-lease owner | `protected_active_workflow_owner` |
+| Instance identity is launcher-minted (`inst-…`) | `untrusted_identity_provenance` |
+| Row's `host_id` matches the host running retirement | `host_binding_unproven` |
+| Boot identity is known on both sides | `boot_identity_unknown` |
+| The pid number is not occupied by a different incarnation | `pid_reuse_detected` |
+
+Eligible rows carry `eligible_stale_orphan`.
+
+Three properties are worth stating explicitly:
+
+* **`pid_alive` can only withdraw liveness, never grant it**
+  (`WorkerRegistry.is_live`, #948 AC7). A heartbeat-lapsed but still-running
+  process therefore classifies as `stale` in the snapshot, yet #980's added
+  `pid_alive is False` requirement preserves it. An unprobeable PID (`None`)
+  also fails closed.
+* **Affirmative identity proof is required (review 657 B2).** An earlier
+  revision required only that the pre-existing registry columns were non-null —
+  which a legacy `legacy-pid-…` row satisfies trivially, so a row that proved
+  nothing about *which* process it described was retireable. Retirement now
+  needs both halves of a positive proof:
+  * **Attribution** — a launcher-minted `inst-…` `client_instance_id`, so the
+    row is known to belong to one specific application launch rather than
+    having been inferred from pid proximity.
+  * **Fencing** — `host_id`, `boot_id`, and `process_start_time`, which turn a
+    bare pid into a statement about one process: which machine it ran on, which
+    boot of that machine, and which incarnation of that pid number.
+
+  **Consequence, stated plainly:** registrations written before these columns
+  existed, and any row on a legacy instance identity, are preserved
+  *permanently*. They are retired only after their worker re-registers under a
+  trusted identity — never on weaker evidence. That the alternative would leave
+  legacy rows outstanding indefinitely is not a reason to relax the proof.
+* **A live pid is an absolute block.** Even across a boot boundary, where the
+  number provably cannot belong to the registered process, an occupied pid
+  preserves the row rather than being argued away by the fencing proof.
+
+Multiple processes belonging to one legitimate worker cohort are not treated as
+multiple independent workers: the fleet model from #948/#978 is preserved
+unchanged, and sharing a role or profile is never a duplicate.
+
+## Tools
+
+### `gitea_plan_stale_worker_retirement`
+
+Read-only. Controller and reconciler only.
+
+| Parameter | Meaning |
+| --- | --- |
+| `remote` | `dadeschools` or `prgs` |
+| `host`, `org`, `repo` | Optional overrides (audit context) |
+| `canonical_repository` | Expected repository binding; defaults to the process root |
+
+Returns `registry_fingerprint`, `candidate_fingerprint`,
+`candidate_worker_identities`, per-worker `candidates` and `preserved` entries
+(each with `reason_code`, `detail`, and structured `evidence`),
+`preserved_reason_counts`, `assessed_count`, `candidate_count`,
+`preserved_count`, and `protected_active_workflow_owners`.
+`mutation_performed` is always `false` and `read_only` is always `true`.
+
+Planning is deterministic: the same authoritative registry contents produce the
+same plan and the same tokens regardless of when they are observed.
+
+### `gitea_apply_stale_worker_retirement`
+
+Mutating. Controller and reconciler only.
+
+| Parameter | Meaning |
+| --- | --- |
+| `registry_fingerprint` | The exact stable token the plan returned |
+| `candidate_fingerprint` | The exact candidate-set token the plan returned |
+| `worker_identities` | The exact candidate identities (list, or JSON / comma-separated string) |
+| `remote`, `host`, `org`, `repo` | As above |
+| `canonical_repository` | Must match the value the plan used |
+
+Before touching the registry, apply fails closed on: profile permission, role
+kind, master parity (`mutation_safe`), stable-runtime mode, capability
+resolution refreshed immediately before mutation, worker-registry availability,
+workflow-lease enumeration failure, and daemon-cohort uniqueness
+(`classify_cohort`).
+
+A matching token is necessary but never sufficient. Inside one
+`BEGIN IMMEDIATE` transaction (`WorkerRegistry.retire_stale_workers`) the
+server:
+
+1. re-reads the authoritative rows;
+2. recomputes `registry_fingerprint` from *those* rows and compares — a mismatch
+   returns `registry_revision_moved` with `retired_count: 0` and no write;
+3. recomputes the eligibility plan from *those* rows — re-reading the active
+   workflow leases rather than reusing the set captured before the transaction
+   opened, so a lease acquired after planning still preserves its worker — and
+   compares `candidate_fingerprint`. A mismatch returns `candidate_set_moved`
+   with `retired_count: 0` and no write; a lease-enumeration failure raises and
+   rolls the transaction back;
+4. revalidates every requested identity against that fresh plan;
+5. retires each survivor with a guarded `UPDATE` that additionally asserts
+   `status`, `last_heartbeat_at`, `generation_id`, `session_id`,
+   `fencing_epoch`, and `pid` are unchanged. A guard that matches no row
+   preserves the worker with `row_changed_since_plan`.
+
+There is no window between a safety check and its matching write, so a worker
+that comes back to life, changes ownership, or is retired concurrently cannot be
+removed on the strength of a stale observation. Any exception — including a
+commit failure — rolls the whole transaction back and returns
+`transaction_failed` with `success: false`, `retired_count: 0`, and
+`mutation_performed: false`; a partial write can never be reported as success.
+
+### Outcomes
+
+| `outcome` | Meaning | `mutation_performed` |
+| --- | --- | --- |
+| `planned` | Dry-run result | `false` |
+| `applied` | Transaction ran; see `retired` / `preserved` | `true` only if something was retired |
+| `registry_revision_moved` | Registry changed between plan and apply | `false` |
+| `candidate_set_moved` | Eligibility verdict changed between plan and apply | `false` |
+| `already_retired` | Every requested row is already retired (idempotent replay) | `false` |
+| `nothing_requested` | Empty target list | `false` |
+| `transaction_failed` | Rolled back; nothing retired | `false` |
+
+## What retirement does to the fleet snapshot
+
+A retired row keeps its history: `status` moves to `retired` and `retired_at`,
+`retired_by`, `retirement_reason` are recorded. Nothing is deleted. Because
+`retired` is not `active`, the #978 snapshot counts the row as **historical**,
+not stale, so `stale_worker_count` falls and historical rows never make the live
+fleet unsafe by themselves.
+
+**Retirement does not repair untrusted live identity.** Live workers registered
+under legacy `pid-`/`proc-` instance identities are preserved untouched and
+keep their `legacy_incomplete_identity` blockers. Retiring every stale row can
+therefore legitimately produce:
+
+* `stale_worker_count: 0`
+* residual live `legacy_incomplete_identity` blockers
+* `live_fleet_safe: false`
+
+That is a truthful result, and the `post_apply` block reports the remaining
+blockers rather than claiming the fleet became safe. Trusted
+`client_instance_id` propagation through launchers is a separate enrollment
+problem.
+
+## Permissions
+
+Plan and apply are authorized differently, and deliberately so (review 657 B1).
+
+| | Plan | Apply |
+| --- | --- | --- |
+| Capability | `gitea.read` | `gitea.worker_registry.retire` |
+| Nature | observational; opens no transaction, writes nothing | mutation |
+| Roles | `controller`, `reconciler` | `controller`, `reconciler` |
+
+An earlier revision authorized apply with `gitea.read` alone, reasoning that
+the mutation lands in the local control-plane registry rather than in Gitea.
+That the write is local makes it **no less a mutation**: sharing an
+observational permission class with plan meant any profile that could *look*
+could also *destroy*. Apply now requires its own capability.
+
+* **Denied:** author, reviewer, merger, and every ordinary read-only profile —
+  they lack the capability, so they fail closed on the permission itself rather
+  than on the role check alone. The role restriction remains as defence in
+  depth: a profile mistakenly granted the capability still cannot reach apply
+  from an author, reviewer, or merger role.
+* The capability is checked at entry **and** re-resolved immediately before the
+  registry mutation, so a profile change mid-call cannot be outrun.
+* **No new Gitea write permission is introduced.**
+  `gitea.worker_registry.retire` authorizes exactly one local control-plane
+  transition (`worker_registrations.status -> retired`) and grants no branch,
+  issue, PR, review, merge, or restart authority. No author permission is
+  broadened.
+* The fleet snapshot remains observational: nothing here turns it into a gate on
+  ordinary author work.
+
+### Operator step
+
+No profile holds `gitea.worker_registry.retire` by default, so apply is inert
+until an operator adds it to the `allowed_operations` of the controller or
+reconciler profile in `profiles.json`. Removing it again immediately and
+completely revokes apply, while leaving plan and every other capability
+untouched. That grant is a configuration change and is outside the scope of the
+code that implements this feature.
+
+## External-state fencing
+
+`BEGIN IMMEDIATE` locks the worker registry and nothing else, so two inputs the
+decision depends on sit outside the transaction's isolation domain: the
+workflow-lease table in a separate control-plane database, and OS process
+liveness. Re-reading them once during revalidation is not sufficient — the
+per-target loop runs afterwards, so a lease acquired (or a pid revived) after
+revalidation but before a given row's `UPDATE` would go unnoticed, and the
+registry-column guard cannot catch it because no registry column changed.
+
+Two mechanisms close that window, both applied per target immediately before
+its own write:
+
+* **`external_fence_fn`** — a version token over active leases
+  (`external_state_fingerprint`), captured inside the transaction *before* the
+  authoritative read and re-compared before every guarded `UPDATE`. Any movement
+  raises, rolling back the whole transaction: once the world has changed, every
+  remaining per-row decision was computed against a world that no longer exists.
+  An unreadable lease store raises rather than returning a token, because
+  "unreadable" must not silently compare equal to "unchanged".
+* **`liveness_fn`** — a re-probe of process liveness and fencing identity that
+  must affirmatively re-establish that this exact process is gone. It compares
+  `process_start_time`, so a pid number reused since the plan is refused rather
+  than accepted.
+
+A caller that supplies no `liveness_fn` retires nothing
+(`liveness_reprobe_unavailable`) rather than proceeding unfenced. The guarded
+`UPDATE` additionally asserts `host_id`, `boot_id`, and `process_start_time` are
+unchanged, and all three participate in the CAS token, so fencing movement
+alone is enough to abort.
+
+## Non-goals
+
+* Killing or restarting processes.
+* Editing session files or configuration.
+* Direct database cleanup outside the sanctioned transaction.
+* Retiring live workers.
+* Backfilling trusted identity for legacy workers.
+* Rewriting worker ownership.
+* Cleaning unrelated workflow-lease or issue-claim registries.

gitea_config.py

@@ -1180,6 +1180,10 @@ RECOGNIZED_GITEA_ENV_KEYS = frozenset({
     "GITEA_SERVER_PROVENANCE",
     "GITEA_AUTHOR_WORKTREE",
     "GITEA_ACTIVE_WORKTREE",
+    "GITEA_REVIEWER_WORKTREE",
+    "GITEA_MERGER_WORKTREE",
+    "GITEA_CANONICAL_REPOSITORY_ROOT",
+    "GITEA_MCP_SESSION_STATE_TTL_HOURS",
     "GITEA_DISABLE_KEYCHAIN",
     "GITEA_CONTROL_PLANE_DB",
     "GITEA_DB_PATH",
@@ -1189,6 +1193,31 @@ RECOGNIZED_GITEA_ENV_KEYS = frozenset({
     "GITEA_IRRECOVERABLE_HMAC_SECRET",
     "GITEA_FORCE_MCP_RUNTIME_CHECK",
     "GITEA_FORCE_CLIENT_MANAGED",
+    # #975: the client-identity inputs the server actually consumes at startup
+    # (CLIENT_NAME_ENV / CLIENT_INSTANCE_ENV / CLIENT_SESSION_ENV in
+    # gitea_mcp_server). Production read them while this allowlist omitted them,
+    # so the peer-env scan classified them as unsupported overrides and the
+    # capability resolver refused every mutation fleet-wide. Named individually
+    # on purpose: no prefix is added, so an unrecognised GITEA_* override is
+    # still refused exactly as it was before.
+    "GITEA_MCP_CLIENT",
+    "GITEA_MCP_CLIENT_INSTANCE",
+    "GITEA_MCP_CLIENT_SESSION",
+    # #978: operator-approved fleet enrollment identity shared by one launch.
+    "GITEA_MCP_FLEET_RUN_ID",
+    "GITEA_MCP_PROCESS_IDENTITY",
+    # #978 B1/B2: launcher-sealed provenance + peer-visible worker identity so
+    # the instance-aware mutation gate can distinguish two legitimate
+    # application instances that share a profile from a true duplicate worker.
+    "GITEA_MCP_INSTANCE_PROVENANCE",
+    "GITEA_MCP_WORKER_IDENTITY",
+    "GITEA_MCP_GENERATION_ID",
+    # #975 review 652 B1: production also consumes HEARTBEAT_INTERVAL_ENV from
+    # mcp_worker_identity via gitea_mcp_server._start_worker_heartbeat. Omitting
+    # it reproduced the same unsupported-env → runtime_reconnect_required
+    # failure mode for the documented operator override. Named individually;
+    # no GITEA_* / GITEA_WORKER_* prefix is added.
+    "GITEA_WORKER_HEARTBEAT_INTERVAL_SECONDS",
 })
 
 RECOGNIZED_GITEA_ENV_PREFIXES = (
@@ -1216,24 +1245,70 @@ def get_unconsumed_gitea_env_overrides(env=None) -> dict[str, str]:
     return unconsumed
 
 
-def launcher_entry(profile_name, config_path=None):
+def launcher_entry(
+    profile_name,
+    config_path=None,
+    *,
+    client_type=None,
+    client_instance_id=None,
+    fleet_run_id=None,
+    session_id=None,
+    launch_nonce=None,
+):
     """Return a thin MCP launcher entry for *profile_name*.
 
-    Contains command/args and the GITEA_MCP_* / GITEA_CLIENT_MANAGED env vars — never a token
-    or password. Suitable for Claude / Gemini / Codex ``mcpServers`` blocks.
+    Contains command/args and the GITEA_MCP_* / GITEA_CLIENT_MANAGED env vars —
+    never a token or password. Suitable for Claude / Gemini / Codex
+    ``mcpServers`` blocks.
+
+    #978 B1: production launches mint (or reuse) a trusted
+    ``GITEA_MCP_CLIENT_INSTANCE`` so workers never fall back to the legacy
+    placeholder identity on the real serve path. Pass *client_instance_id* to
+    resume the same launch; omit it for a fresh launch (new ID).
     """
-    command, args = server_command()
-    return {
-        "gitea-tools": {
-            "command": command,
-            "args": args,
-            "env": {
-                "GITEA_MCP_CONFIG": config_path or DEFAULT_CONFIG_PATH,
-                "GITEA_MCP_PROFILE": profile_name,
-                "GITEA_CLIENT_MANAGED": "1",
-            },
-        }
-    }
+    import mcp_application_launcher as app_launcher
+
+    built = app_launcher.launcher_entry_for_profile(
+        profile_name,
+        client_type=client_type or "unknown",
+        config_path=config_path or DEFAULT_CONFIG_PATH,
+        client_instance_id=client_instance_id,
+        fleet_run_id=fleet_run_id,
+        session_id=session_id,
+        launch_nonce=launch_nonce,
+        server_key="gitea-tools",
+    )
+    # Public shape stays {server_key: {command, args, env}} for existing callers.
+    return {"gitea-tools": built["gitea-tools"]}
+
+
+def multi_namespace_launcher_entries(
+    profile_by_namespace,
+    *,
+    client_type,
+    config_path=None,
+    client_instance_id=None,
+    fleet_run_id=None,
+    session_id=None,
+    launch_nonce=None,
+):
+    """Build production mcpServers for all five namespaces of one application launch.
+
+    One shared trusted ``client_instance_id`` is minted (or reused) and
+    propagated to every namespace worker env. See
+    :func:`mcp_application_launcher.build_application_mcp_servers`.
+    """
+    import mcp_application_launcher as app_launcher
+
+    return app_launcher.build_application_mcp_servers(
+        profile_by_namespace,
+        client_type=client_type,
+        config_path=config_path or DEFAULT_CONFIG_PATH,
+        client_instance_id=client_instance_id,
+        fleet_run_id=fleet_run_id,
+        session_id=session_id,
+        launch_nonce=launch_nonce,
+    )
 
 
 

mcp_application_launcher.py

@@ -0,0 +1,393 @@
+"""Production application launcher for multi-namespace MCP fleets (#978 B1).
+
+One real LLM application launch mints exactly one trusted
+``client_instance_id`` and propagates it to every Gitea MCP namespace worker
+started for that launch. Workers never invent a trusted instance identity from
+PID proximity, timestamps, or ordinary untrusted environment values.
+
+This module is the production serve-path authority for instance identity.
+Tests and fixtures may call the same functions, but production registration
+receives the identity from the env this launcher builds — not from a hand-set
+test-only helper that bypasses it.
+"""
+
+from __future__ import annotations
+
+import os
+import secrets
+from typing import Any, Mapping
+
+import mcp_fleet_snapshot as fleet
+import mcp_worker_identity as mwi
+
+#: Canonical MCP server names for the five role namespaces.
+NAMESPACE_SERVER_NAMES: tuple[str, ...] = (
+    "gitea-author",
+    "gitea-reviewer",
+    "gitea-merger",
+    "gitea-controller",
+    "gitea-reconciler",
+)
+
+#: Map MCP server name → role/namespace kind.
+SERVER_TO_NAMESPACE: dict[str, str] = {
+    "gitea-author": "author",
+    "gitea-reviewer": "reviewer",
+    "gitea-merger": "merger",
+    "gitea-controller": "controller",
+    "gitea-reconciler": "reconciler",
+}
+
+SANCTIONED_NAMESPACES: tuple[str, ...] = (
+    "author",
+    "reviewer",
+    "merger",
+    "controller",
+    "reconciler",
+)
+
+# Env the launcher may set on every worker of one application launch.
+CLIENT_NAME_ENV = "GITEA_MCP_CLIENT"
+CLIENT_INSTANCE_ENV = fleet.CLIENT_INSTANCE_ENV
+FLEET_RUN_ENV = fleet.FLEET_RUN_ENV
+CLIENT_SESSION_ENV = "GITEA_MCP_CLIENT_SESSION"
+CLIENT_MANAGED_ENV = "GITEA_CLIENT_MANAGED"
+PROFILE_ENV = "GITEA_MCP_PROFILE"
+CONFIG_ENV = "GITEA_MCP_CONFIG"
+INSTANCE_PROVENANCE_ENV = "GITEA_MCP_INSTANCE_PROVENANCE"
+WORKER_IDENTITY_ENV = "GITEA_MCP_WORKER_IDENTITY"
+GENERATION_ID_ENV = "GITEA_MCP_GENERATION_ID"
+
+# Marker the trusted launcher alone writes; ordinary user env without this
+# marker is never classified as launcher-trusted provenance.
+LAUNCHER_PROVENANCE_VALUE = fleet.INSTANCE_ID_PROVENANCE_TRUSTED
+
+
+def mint_application_launch(
+    client_type: str | None,
+    *,
+    launch_nonce: str | None = None,
+    fleet_run_id: str | None = None,
+    session_id: str | None = None,
+    now=None,
+) -> dict[str, Any]:
+    """Mint one trusted application-instance identity for a production launch.
+
+    Called exactly once per real application launch. The returned
+    ``client_instance_id`` is injected into every namespace worker environment
+    for that launch. A second call (separate launch) yields a different ID.
+    """
+    client = mwi.normalize_client_name(client_type)
+    instance_id = fleet.generate_client_instance_id(
+        client, launch_nonce=launch_nonce, now=now
+    )
+    assessment = fleet.assess_instance_identity(instance_id)
+    if not assessment["trusted"]:
+        # generate_client_instance_id always produces a trusted format; fail
+        # closed if that invariant ever breaks rather than shipping untrusted.
+        raise RuntimeError(
+            f"launcher produced untrusted client_instance_id {instance_id!r}: "
+            f"{assessment.get('reasons')}"
+        )
+    session = (session_id or "").strip() or f"launch-{secrets.token_hex(12)}"
+    return {
+        "client_type": client,
+        "client_instance_id": instance_id,
+        "instance_id_provenance": LAUNCHER_PROVENANCE_VALUE,
+        "instance_identity_trusted": True,
+        "fleet_run_id": (fleet_run_id or "").strip() or None,
+        "session_id": session,
+        "namespaces": list(SANCTIONED_NAMESPACES),
+        "namespace_server_names": list(NAMESPACE_SERVER_NAMES),
+    }
+
+
+def namespace_worker_env(
+    *,
+    profile_name: str,
+    client_type: str | None,
+    client_instance_id: str,
+    config_path: str | None = None,
+    fleet_run_id: str | None = None,
+    session_id: str | None = None,
+    extra_env: Mapping[str, str] | None = None,
+) -> dict[str, str]:
+    """Build the environment for one namespace worker of a trusted launch.
+
+    Never invents a client_instance_id. The caller must supply the launch-minted
+    identity so all five workers receive the same value.
+    """
+    assessment = fleet.assess_instance_identity(client_instance_id)
+    if not assessment["trusted"]:
+        raise ValueError(
+            "namespace_worker_env refuses untrusted client_instance_id "
+            f"{client_instance_id!r}: {assessment.get('reasons')}"
+        )
+    client = mwi.normalize_client_name(client_type)
+    env: dict[str, str] = {
+        PROFILE_ENV: str(profile_name),
+        CLIENT_MANAGED_ENV: "1",
+        "GITEA_MCP_CLIENT": client,
+        CLIENT_INSTANCE_ENV: assessment["client_instance_id"],
+        INSTANCE_PROVENANCE_ENV: LAUNCHER_PROVENANCE_VALUE,
+    }
+    if config_path:
+        env[CONFIG_ENV] = str(config_path)
+    if fleet_run_id:
+        env[FLEET_RUN_ENV] = str(fleet_run_id)
+    if session_id:
+        env[CLIENT_SESSION_ENV] = str(session_id)
+    if extra_env:
+        # Never let untrusted callers override the trusted instance keys.
+        protected = {
+            CLIENT_INSTANCE_ENV,
+            INSTANCE_PROVENANCE_ENV,
+            "GITEA_MCP_CLIENT",
+            CLIENT_MANAGED_ENV,
+        }
+        for key, value in extra_env.items():
+            if key in protected:
+                continue
+            env[str(key)] = str(value)
+    return env
+
+
+def build_application_mcp_servers(
+    profile_by_namespace: Mapping[str, str],
+    *,
+    client_type: str | None,
+    config_path: str | None = None,
+    client_instance_id: str | None = None,
+    fleet_run_id: str | None = None,
+    session_id: str | None = None,
+    launch_nonce: str | None = None,
+    command: str | None = None,
+    args: list[str] | None = None,
+    now=None,
+) -> dict[str, Any]:
+    """Build a production ``mcpServers`` map for one application launch.
+
+    Mints one ``client_instance_id`` when *client_instance_id* is omitted (fresh
+    launch). When the caller supplies a previously minted trusted ID (resume of
+    the same launch / config rewrite), that ID is reused so reconnect keeps
+    attribution. A full new application restart omits the ID and receives a
+    fresh mint.
+
+    Every namespace server entry receives the **same** instance ID. Separate
+    calls with no supplied ID receive distinct IDs.
+    """
+    missing = [
+        ns for ns in SANCTIONED_NAMESPACES if not profile_by_namespace.get(ns)
+    ]
+    if missing:
+        raise ValueError(
+            "build_application_mcp_servers requires a profile for every "
+            f"sanctioned namespace; missing: {missing}"
+        )
+
+    if client_instance_id is None:
+        launch = mint_application_launch(
+            client_type,
+            launch_nonce=launch_nonce,
+            fleet_run_id=fleet_run_id,
+            session_id=session_id,
+            now=now,
+        )
+    else:
+        assessment = fleet.assess_instance_identity(client_instance_id)
+        if not assessment["trusted"]:
+            raise ValueError(
+                "refusing to propagate untrusted client_instance_id "
+                f"{client_instance_id!r} into production launch envs: "
+                f"{assessment.get('reasons')}"
+            )
+        launch = {
+            "client_type": mwi.normalize_client_name(client_type),
+            "client_instance_id": assessment["client_instance_id"],
+            "instance_id_provenance": LAUNCHER_PROVENANCE_VALUE,
+            "instance_identity_trusted": True,
+            "fleet_run_id": (fleet_run_id or "").strip() or None,
+            "session_id": (session_id or "").strip()
+            or f"launch-{secrets.token_hex(12)}",
+            "namespaces": list(SANCTIONED_NAMESPACES),
+            "namespace_server_names": list(NAMESPACE_SERVER_NAMES),
+        }
+
+    # Resolve command/args from the production server entry when not provided.
+    if command is None or args is None:
+        import gitea_config
+
+        cmd, cmd_args = gitea_config.server_command()
+        command = command or cmd
+        args = args if args is not None else list(cmd_args)
+
+    servers: dict[str, Any] = {}
+    shared_id = launch["client_instance_id"]
+    for namespace in SANCTIONED_NAMESPACES:
+        server_name = f"gitea-{namespace}"
+        profile = profile_by_namespace[namespace]
+        env = namespace_worker_env(
+            profile_name=profile,
+            client_type=launch["client_type"],
+            client_instance_id=shared_id,
+            config_path=config_path,
+            fleet_run_id=launch.get("fleet_run_id"),
+            session_id=launch.get("session_id"),
+        )
+        servers[server_name] = {
+            "command": command,
+            "args": list(args),
+            "env": env,
+        }
+
+    return {
+        "mcpServers": servers,
+        "launch": launch,
+        "client_instance_id": shared_id,
+        "client_type": launch["client_type"],
+        "namespaces": list(SANCTIONED_NAMESPACES),
+        "shared_instance_id_across_namespaces": True,
+        "namespace_count": len(SANCTIONED_NAMESPACES),
+    }
+
+
+def launcher_entry_for_profile(
+    profile_name: str,
+    *,
+    client_type: str | None = None,
+    config_path: str | None = None,
+    client_instance_id: str | None = None,
+    fleet_run_id: str | None = None,
+    session_id: str | None = None,
+    server_key: str = "gitea-tools",
+    launch_nonce: str | None = None,
+    now=None,
+) -> dict[str, Any]:
+    """Thin single-server production launcher entry with trusted instance ID.
+
+    Used when only one namespace is being configured. Still mints (or reuses)
+    a trusted ``client_instance_id`` so production never relies on the legacy
+    placeholder identity for normal launches.
+    """
+    import gitea_config
+
+    if client_instance_id is None:
+        launch = mint_application_launch(
+            client_type or "unknown",
+            launch_nonce=launch_nonce,
+            fleet_run_id=fleet_run_id,
+            session_id=session_id,
+            now=now,
+        )
+        client_instance_id = launch["client_instance_id"]
+        client = launch["client_type"]
+        fleet_run = launch.get("fleet_run_id")
+        session = launch.get("session_id")
+    else:
+        assessment = fleet.assess_instance_identity(client_instance_id)
+        if not assessment["trusted"]:
+            raise ValueError(
+                f"untrusted client_instance_id {client_instance_id!r}"
+            )
+        client = mwi.normalize_client_name(client_type)
+        fleet_run = (fleet_run_id or "").strip() or None
+        session = (session_id or "").strip() or None
+        client_instance_id = assessment["client_instance_id"]
+
+    command, args = gitea_config.server_command()
+    env = namespace_worker_env(
+        profile_name=profile_name,
+        client_type=client,
+        client_instance_id=client_instance_id,
+        config_path=config_path or gitea_config.DEFAULT_CONFIG_PATH,
+        fleet_run_id=fleet_run,
+        session_id=session,
+    )
+    return {
+        server_key: {
+            "command": command,
+            "args": args,
+            "env": env,
+        },
+        "client_instance_id": client_instance_id,
+        "client_type": client,
+    }
+
+
+def collect_instance_ids_from_mcp_servers(
+    mcp_servers: Mapping[str, Any],
+) -> dict[str, Any]:
+    """Inspect a production mcpServers map for shared instance attribution.
+
+    Returns the unique set of client_instance_id values across gitea-* servers
+    and whether all five namespaces share exactly one trusted ID.
+    """
+    ids: list[str] = []
+    by_server: dict[str, str | None] = {}
+    for name in NAMESPACE_SERVER_NAMES:
+        entry = mcp_servers.get(name) or {}
+        env = entry.get("env") or {}
+        raw = (env.get(CLIENT_INSTANCE_ENV) or "").strip() or None
+        by_server[name] = raw
+        if raw:
+            ids.append(raw)
+    unique = sorted(set(ids))
+    trusted = [
+        i
+        for i in unique
+        if fleet.assess_instance_identity(i)["trusted"]
+    ]
+    return {
+        "server_instance_ids": by_server,
+        "unique_instance_ids": unique,
+        "trusted_instance_ids": trusted,
+        "shared_single_trusted_id": (
+            len(unique) == 1
+            and len(trusted) == 1
+            and all(by_server.get(n) == unique[0] for n in NAMESPACE_SERVER_NAMES)
+        ),
+        "namespace_server_count": sum(
+            1 for n in NAMESPACE_SERVER_NAMES if n in mcp_servers
+        ),
+    }
+
+
+def inherit_or_refuse_client_instance(
+    env: Mapping[str, str] | None = None,
+) -> dict[str, Any]:
+    """Resolve instance identity for a worker process at serve time.
+
+    Production workers inherit the launcher-issued ID. They never mint a
+    trusted ID themselves. Missing / legacy / malformed values fail soft into
+    an untrusted assessment so registration can still record a diagnostic row
+    without authorizing multi-instance fleet mutation.
+    """
+    source = dict(env if env is not None else os.environ)
+    raw = (source.get(CLIENT_INSTANCE_ENV) or "").strip() or None
+    provenance_marker = (source.get(INSTANCE_PROVENANCE_ENV) or "").strip()
+    assessment = fleet.assess_instance_identity(raw)
+    # Ordinary user-supplied values without launcher provenance marker are
+    # still format-checked by assess_instance_identity. When the format is
+    # trusted but the launcher marker is absent, keep the ID but note that
+    # provenance is not launcher-sealed (operator hand-set or legacy config).
+    if assessment["trusted"] and provenance_marker != LAUNCHER_PROVENANCE_VALUE:
+        assessment = dict(assessment)
+        assessment["launcher_sealed"] = False
+        assessment["reasons"] = list(assessment.get("reasons") or []) + [
+            f"{INSTANCE_PROVENANCE_ENV} is not {LAUNCHER_PROVENANCE_VALUE!r}; "
+            "identity format is valid but not sealed by the production launcher"
+        ]
+    else:
+        assessment = dict(assessment)
+        assessment["launcher_sealed"] = bool(
+            assessment["trusted"]
+            and provenance_marker == LAUNCHER_PROVENANCE_VALUE
+        )
+    assessment["fleet_run_id"] = (source.get(FLEET_RUN_ENV) or "").strip() or None
+    assessment["session_id"] = (
+        (source.get(CLIENT_SESSION_ENV) or "").strip() or None
+    )
+    assessment["client_type"] = mwi.normalize_client_name(
+        (source.get("GITEA_MCP_CLIENT") or "").strip() or None
+    )
+    return assessment

mcp_fleet_retirement.py

@@ -0,0 +1,776 @@
+"""CAS-protected retirement planning for stale worker registrations (#980).
+
+#978 (merged PR #979) made the fleet observable: every registered namespace
+worker, its instance attribution, its heartbeat freshness, and a structured
+classification. It deliberately stopped there — the snapshot is read-only and
+the control plane still had no sanctioned way to retire registry rows whose
+owning process is conclusively gone.
+
+This module is the *decision layer* for that retirement. It is pure: callers
+supply registry rows, a clock, and a PID probe; nothing here opens SQLite,
+scans process tables, or mutates state. The transactional apply lives in
+:meth:`mcp_worker_identity.WorkerRegistry.retire_stale_workers`, which calls
+back into these same pure functions so plan and apply can never disagree about
+what "the registry looks like" or "which rows are eligible".
+
+Why a separate token
+--------------------
+
+``mcp_fleet_snapshot._consistency_token`` seeds its digest with ``snapshot_at``
+at second precision, so ``registry_revision`` changes on every call even when
+no registry row changed. A compare-and-swap gated on it can never pass — a
+dry-run/apply cycle spanning more than one second aborts unconditionally. That
+token is still useful as an observation stamp, so it is left exactly as it is;
+#980 gets its own :func:`registry_fingerprint`, derived *only* from canonical
+retirement-relevant row content:
+
+* identical registry contents observed at any two times produce the same token,
+* row order never affects the token (serialized rows are sorted),
+* any create/delete/identity/liveness/ownership/registration-state change to a
+  retirement-relevant field changes the token.
+
+Fail-closed posture
+-------------------
+
+A worker is retired only when the control plane *conclusively* establishes it
+is a stale orphan. Missing evidence is never read as permission: an unprobeable
+PID, an unparsable heartbeat, a row that shares identity evidence with a live
+or unprobeable worker, a foreign or absent repository binding, or a worker that
+still owns an active workflow lease all preserve the row.
+
+Trusted launcher identity (``inst-…`` provenance) is deliberately *not* part of
+the conjunction. #980 places trusted ``client_instance_id`` propagation out of
+scope and lists "backfilling trusted identity for legacy workers" as a non-goal;
+requiring it here would preserve every legacy row forever and make the feature
+inert. What *is* required is that the registry fields the conjunction reads are
+actually present — see :data:`REQUIRED_IDENTITY_FIELDS`.
+"""
+
+from __future__ import annotations
+
+import hashlib
+from datetime import datetime
+from typing import Any, Callable, Iterable, Mapping, Sequence
+
+import mcp_fleet_snapshot as fleet
+import mcp_worker_identity as mwi
+
+# --- Outcomes -------------------------------------------------------------
+
+OUTCOME_PLANNED = "planned"
+OUTCOME_APPLIED = "applied"
+OUTCOME_REGISTRY_MOVED = "registry_revision_moved"
+OUTCOME_CANDIDATES_MOVED = "candidate_set_moved"
+OUTCOME_ALREADY_RETIRED = "already_retired"
+OUTCOME_NOTHING_REQUESTED = "nothing_requested"
+
+# --- Reason codes ---------------------------------------------------------
+
+#: The only reason code that authorizes retirement.
+REASON_ELIGIBLE = "eligible_stale_orphan"
+
+REASON_ALREADY_TERMINAL = "already_terminal_registration"
+REASON_AMBIGUOUS_OWNERSHIP = "ambiguous_ownership_state"
+REASON_CONFLICTING_IDENTITY = "conflicting_identity_evidence"
+REASON_FOREIGN_REPOSITORY = "repository_binding_ambiguous"
+REASON_HEARTBEAT_FRESH = "heartbeat_not_expired"
+REASON_INCOMPLETE_IDENTITY = "incomplete_registry_identity"
+REASON_NOT_IN_PLAN = "not_in_current_plan"
+REASON_PID_ALIVE = "pid_alive"
+REASON_PID_UNKNOWN = "pid_liveness_unknown"
+REASON_PROTECTED_OWNER = "protected_active_workflow_owner"
+REASON_ROW_CHANGED = "row_changed_since_plan"
+REASON_ROW_MISSING = "registration_missing"
+REASON_UNPARSABLE_HEARTBEAT = "unparsable_heartbeat"
+REASON_WORKER_LIVE = "worker_live"
+
+# --- #980 review 657 B2: affirmative identity/liveness proof ---------------
+
+#: The registration's instance identity is not launcher-minted (``inst-…``),
+#: so nothing proves which application launch this row belongs to.
+REASON_UNTRUSTED_PROVENANCE = "untrusted_identity_provenance"
+#: The row does not record which host its pid belongs to, or records a
+#: different host than the one probing. A local pid probe cannot speak for a
+#: process on another machine.
+REASON_HOST_UNPROVEN = "host_binding_unproven"
+#: Boot identity is missing on the row or unobtainable here, so a recorded pid
+#: cannot be compared against a live pid at all.
+REASON_BOOT_UNKNOWN = "boot_identity_unknown"
+#: The pid is alive but belongs to a different process incarnation than the one
+#: registered — reported distinctly from a plain live worker.
+REASON_PID_REUSED = "pid_reuse_detected"
+#: Two active registrations claim one client instance within one namespace.
+REASON_INSTANCE_CONFLICT = "client_instance_conflict"
+#: The immediate pre-write re-probe could not re-establish death (#980 B3).
+REASON_LIVENESS_REPROBE = "liveness_reprobe_refused"
+
+#: Instance-identity prefix minted by the trusted launcher. Kept in sync with
+#: ``mcp_fleet_snapshot._TRUSTED_INSTANCE_PREFIX`` through
+#: :func:`mcp_fleet_snapshot.assess_instance_identity`, which stays the single
+#: authority on what "trusted" means — this module never re-implements it.
+TRUSTED_INSTANCE_PREFIX = "inst-"
+
+#: Registry columns that must carry a usable value before the eligibility
+#: conjunction can even be evaluated. Absence is ambiguity, not permission.
+#:
+#: #980 review 657 B2 added the fencing triple. Before it, "complete identity"
+#: meant only that the pre-existing columns were non-null, which a legacy
+#: ``legacy-pid-…`` row satisfies trivially — so a row that proved nothing about
+#: *which* process it described was retireable. The triple is what makes a
+#: recorded pid interpretable: which machine it ran on, which boot of that
+#: machine, and which incarnation of that pid number. A registration written
+#: before these columns existed carries NULL and is therefore preserved
+#: permanently, which is the intended fail-closed outcome.
+REQUIRED_IDENTITY_FIELDS: tuple[str, ...] = (
+    "worker_identity",
+    "client_instance_id",
+    "session_id",
+    "generation_id",
+    "status",
+    "started_at",
+    "last_heartbeat_at",
+    "heartbeat_ttl_seconds",
+    "pid",
+    "host_id",
+    "boot_id",
+    "process_start_time",
+)
+
+#: Canonical retirement-relevant content. Ordering here is fixed and part of
+#: the token contract; adding a field changes every fingerprint, so a change
+#: here is a deliberate contract revision.
+#:
+#: Deliberately excluded: ``token_fingerprint`` (credential-adjacent, never a
+#: retirement input), the four ``*_revision`` columns (revision drift is an
+#: independent restart concern and is not part of the eligibility conjunction),
+#: and the ``retired_*`` bookkeeping columns this feature adds.
+FINGERPRINT_FIELDS: tuple[str, ...] = (
+    "worker_identity",
+    "client_name",
+    "client_instance_id",
+    "session_id",
+    "generation_id",
+    "role",
+    "profile",
+    "namespace",
+    "remote",
+    "repository_binding",
+    "pid",
+    "process_identity",
+    "transport",
+    "started_at",
+    "last_heartbeat_at",
+    "heartbeat_ttl_seconds",
+    "fencing_epoch",
+    "status",
+    "fleet_run_id",
+    "authenticated_account",
+    "instance_id_provenance",
+    # #980 review 657 B3: fencing evidence is a retirement input, so moving it
+    # must move the CAS token. Without these, a row whose host, boot, or
+    # process incarnation changed would hash identically to the row the plan
+    # approved.
+    "host_id",
+    "boot_id",
+    "process_start_time",
+)
+
+_FINGERPRINT_VERSION = "registryfp-v1"
+_CANDIDATE_VERSION = "candidatefp-v1"
+_UNIT = "\x1f"
+_RECORD = "\x1e"
+
+
+def _canon(value: Any) -> str:
+    """Stable text for one field value, independent of Python/SQLite typing.
+
+    ``900`` and ``900.0`` are the same TTL and must hash the same; a value that
+    round-trips through SQLite as REAL must not produce a different token than
+    the same value supplied by a caller as ``int``.
+    """
+    if value is None:
+        return ""
+    if isinstance(value, bool):
+        return "true" if value else "false"
+    if isinstance(value, float):
+        if value != value or value in (float("inf"), float("-inf")):
+            return repr(value)
+        if value.is_integer():
+            return str(int(value))
+        return repr(value)
+    if isinstance(value, int):
+        return str(value)
+    return str(value)
+
+
+def _serialize_row(row: Mapping[str, Any]) -> str:
+    return _UNIT.join(f"{name}={_canon(row.get(name))}" for name in FINGERPRINT_FIELDS)
+
+
+def _digest(version: str, serialized: Sequence[str], prefix: str) -> str:
+    ordered = sorted(serialized)
+    material = _RECORD.join([version, str(len(ordered)), *ordered])
+    return f"{prefix}-{hashlib.sha256(material.encode('utf-8')).hexdigest()[:32]}"
+
+
+def registry_fingerprint(rows: Iterable[Mapping[str, Any]]) -> str:
+    """Content-derived compare-and-swap token for the worker registry (#980).
+
+    Derived exclusively from :data:`FINGERPRINT_FIELDS` across every row. It
+    contains no ``snapshot_at``, wall-clock, request, or report-generation
+    time, so two observations of an unchanged registry always agree, and the
+    serialized rows are sorted so iteration order cannot perturb the digest.
+    """
+    return _digest(
+        _FINGERPRINT_VERSION,
+        [_serialize_row(row) for row in rows],
+        "registryfp",
+    )
+
+
+def candidate_fingerprint(candidate_rows: Iterable[Mapping[str, Any]]) -> str:
+    """Exact-candidate-set token over the selected rows' canonical content.
+
+    A matching :func:`registry_fingerprint` already implies these rows are
+    unchanged; this second token additionally pins *which* rows the operator
+    approved, so an apply can never widen or narrow the approved set.
+    """
+    return _digest(
+        _CANDIDATE_VERSION,
+        [_serialize_row(row) for row in candidate_rows],
+        "candidatefp",
+    )
+
+
+def _probe_pid(
+    pid: Any, pid_alive_probe: Callable[[int | None], bool | None] | None
+) -> bool | None:
+    if pid_alive_probe is None or pid is None:
+        return None
+    try:
+        result = pid_alive_probe(pid)
+    except Exception:
+        return None
+    return None if result is None else bool(result)
+
+
+def _reuse_detected(
+    row: Mapping[str, Any],
+    live_start_time: str | None,
+    current_host_id: str | None,
+    current_boot_id: str | None,
+) -> bool:
+    """Is the pid occupied by a *different* incarnation than the one recorded?
+
+    Only meaningful when the recorded pid is comparable to the live one — same
+    machine, same boot. Across hosts or boots the number is unrelated by
+    construction and reuse is not the interesting question.
+    """
+    recorded_host = (row.get("host_id") or "").strip()
+    recorded_boot = (row.get("boot_id") or "").strip()
+    if not current_host_id or recorded_host != current_host_id:
+        return False
+    if not current_boot_id or recorded_boot != current_boot_id:
+        return False
+    recorded_start = (row.get("process_start_time") or "").strip()
+    return bool(live_start_time) and live_start_time != recorded_start
+
+
+def _instance_key(row: Mapping[str, Any]) -> tuple[str, str] | None:
+    """The (instance, namespace) pair #978 requires to be unique among live rows."""
+    instance = _canon(row.get("client_instance_id"))
+    namespace = _canon(row.get("namespace"))
+    if not instance or not namespace:
+        return None
+    return (instance, namespace)
+
+
+def _probe_start_time(
+    pid: Any, start_time_probe: Callable[[Any], str | None] | None
+) -> str | None:
+    if start_time_probe is None or pid is None:
+        return None
+    try:
+        return start_time_probe(pid)
+    except Exception:
+        return None
+
+
+def _evidence(
+    row: Mapping[str, Any],
+    snapshot_row: Mapping[str, Any],
+    pid_alive: bool | None,
+) -> dict[str, Any]:
+    liveness = snapshot_row.get("liveness") or {}
+    return {
+        "worker_identity": row.get("worker_identity"),
+        "client_type": snapshot_row.get("client_type"),
+        "client_instance_id": row.get("client_instance_id"),
+        "fleet_run_id": row.get("fleet_run_id"),
+        "namespace": row.get("namespace"),
+        "profile": row.get("profile"),
+        "declared_role": row.get("role"),
+        "session_id": row.get("session_id"),
+        "generation_id": row.get("generation_id"),
+        "fencing_epoch": row.get("fencing_epoch"),
+        "process_identity": snapshot_row.get("process_identity"),
+        "pid": row.get("pid"),
+        "pid_alive": pid_alive,
+        "host_id": row.get("host_id"),
+        "boot_id": row.get("boot_id"),
+        "process_start_time": row.get("process_start_time"),
+        "repository_binding": row.get("repository_binding"),
+        "foreign_repository": bool(snapshot_row.get("foreign_repository")),
+        "status": row.get("status"),
+        "started_at": row.get("started_at"),
+        "last_heartbeat_at": row.get("last_heartbeat_at"),
+        "heartbeat_ttl_seconds": row.get("heartbeat_ttl_seconds"),
+        "heartbeat_age_seconds": liveness.get("heartbeat_age_seconds"),
+        "heartbeat_fresh": liveness.get("heartbeat_fresh"),
+        "live": bool(snapshot_row.get("live")),
+        "ownership_state": snapshot_row.get("ownership_state"),
+        "instance_id_provenance": snapshot_row.get("instance_id_provenance"),
+        "instance_identity_trusted": bool(
+            snapshot_row.get("instance_identity_trusted")
+        ),
+    }
+
+
+def _conflict_keys(
+    row: Mapping[str, Any], snapshot_row: Mapping[str, Any]
+) -> list[tuple[str, str]]:
+    keys: list[tuple[str, str]] = []
+    for name, value in (
+        ("session_id", row.get("session_id")),
+        ("generation_id", row.get("generation_id")),
+        ("process_identity", snapshot_row.get("process_identity")),
+        ("pid", row.get("pid")),
+    ):
+        text = _canon(value)
+        if text:
+            keys.append((name, text))
+    return keys
+
+
+def external_state_fingerprint(
+    leases: Iterable[Mapping[str, Any]],
+    *,
+    liveness: Iterable[tuple[Any, Any]] = (),
+) -> str:
+    """Version token over the external state a retirement decision consumed.
+
+    #980 review 657 B3: ``BEGIN IMMEDIATE`` on the worker registry does not
+    cover the control-plane lease table or the OS process table, so those
+    inputs can move while the transaction is open. This token lets the
+    transaction detect that movement: it is captured before the authoritative
+    read and re-compared immediately before every guarded write, and any
+    difference aborts rather than retiring against evidence that has changed.
+
+    Only ownership-relevant lease fields participate, so unrelated churn (a
+    heartbeat timestamp advancing on an unrelated lease) does not cause
+    spurious aborts, while an acquire, release, or owner change always does.
+    """
+    lease_units: list[str] = []
+    for lease in leases:
+        lease_units.append(
+            _UNIT.join(
+                f"{name}={_canon(lease.get(name))}"
+                for name in (
+                    "lease_id",
+                    "role",
+                    "target",
+                    "status",
+                    "session_id",
+                    "owner_session_id",
+                    "owner_pid",
+                    "session_pid",
+                    "generation",
+                )
+            )
+        )
+    for pid, alive in liveness:
+        lease_units.append(f"liveness{_UNIT}pid={_canon(pid)}{_UNIT}alive={_canon(alive)}")
+    return _digest("externalfp-v1", lease_units, "externalfp")
+
+
+def assess_retirement_identity_proof(
+    row: Mapping[str, Any],
+    snapshot_row: Mapping[str, Any],
+    *,
+    current_host_id: str | None,
+    current_boot_id: str | None,
+    live_start_time: str | None,
+    pid_alive: bool | None,
+) -> tuple[str, str] | None:
+    """Affirmative proof that this row names one specific, now-dead process.
+
+    Returns ``None`` when the proof holds, or ``(reason_code, detail)`` naming
+    the first thing that could not be established. #980 review 657 B2: absence
+    of evidence is never read as staleness, so every branch here refuses on
+    *missing* information exactly as firmly as on contradictory information.
+
+    The proof has two independent halves and needs both:
+
+    * **Attribution** — a launcher-minted ``inst-…`` instance identity, so the
+      row is known to belong to one specific application launch rather than
+      having been inferred from pid proximity.
+    * **Fencing** — the row's host matches the host doing the probing, boot
+      identity is known on both sides, and the recorded process incarnation
+      agrees with whatever currently occupies that pid number.
+
+    Requiring trusted attribution means pre-#978 ``legacy-pid-…`` rows are
+    preserved permanently. That is deliberate. The reviewer specifically
+    rejected the argument that legacy rows "would remain forever" as grounds
+    for a weaker proof, and #980 lists backfilling trusted identity for legacy
+    workers as a non-goal — so those rows are retired only after their worker
+    re-registers under a trusted identity, never on weaker evidence.
+    """
+    if not snapshot_row.get("instance_identity_trusted"):
+        return (
+            REASON_UNTRUSTED_PROVENANCE,
+            "client_instance_id "
+            f"{row.get('client_instance_id')!r} is not launcher-minted "
+            f"({snapshot_row.get('instance_id_provenance')!r}); nothing proves "
+            "which application launch this registration belongs to",
+        )
+
+    recorded_host = (row.get("host_id") or "").strip()
+    if not current_host_id:
+        return (
+            REASON_HOST_UNPROVEN,
+            "this process cannot establish its own host identity, so a local "
+            "pid probe cannot be attributed to any machine",
+        )
+    if recorded_host != current_host_id:
+        return (
+            REASON_HOST_UNPROVEN,
+            f"registration is bound to host {recorded_host!r} but retirement is "
+            f"running on {current_host_id!r}; a local pid probe says nothing "
+            "about a process on another machine",
+        )
+
+    recorded_boot = (row.get("boot_id") or "").strip()
+    if not current_boot_id:
+        return (
+            REASON_BOOT_UNKNOWN,
+            "the current boot identity could not be determined, so a recorded "
+            "pid cannot be compared against a live pid",
+        )
+
+    recorded_start = (row.get("process_start_time") or "").strip()
+
+    if recorded_boot != current_boot_id:
+        # A different boot is the strongest possible death evidence: every pid
+        # from a previous boot is gone, and pid numbers restart, so whatever
+        # occupies this number now is unrelated by construction.
+        return None
+
+    # Same boot: the pid number is directly comparable, so the recorded
+    # incarnation must still agree with whatever holds that number.
+    if pid_alive and live_start_time and live_start_time != recorded_start:
+        return (
+            REASON_PID_REUSED,
+            f"pid {row.get('pid')!r} is alive but started at "
+            f"{live_start_time!r}, not the registered {recorded_start!r}; the "
+            "number was reused by an unrelated process and this registration's "
+            "own liveness is therefore unproven",
+        )
+    if pid_alive:
+        return (
+            REASON_PID_ALIVE,
+            f"recorded pid {row.get('pid')!r} is still running on this host and "
+            "boot",
+        )
+    return None
+
+
+def _missing_identity_fields(row: Mapping[str, Any]) -> list[str]:
+    missing: list[str] = []
+    for name in REQUIRED_IDENTITY_FIELDS:
+        value = row.get(name)
+        if value is None or (isinstance(value, str) and not value.strip()):
+            missing.append(name)
+    return missing
+
+
+def plan_stale_worker_retirement(
+    rows: Iterable[Mapping[str, Any]],
+    *,
+    now: datetime | None = None,
+    pid_alive_probe: Callable[[int | None], bool | None] | None = None,
+    canonical_repository: str | None = None,
+    protected_worker_identities: Iterable[str] | None = None,
+    protected_session_ids: Iterable[str] | None = None,
+    protected_pids: Iterable[Any] | None = None,
+    current_host_id: str | None = None,
+    current_boot_id: str | None = None,
+    start_time_probe: Callable[[Any], str | None] | None = None,
+) -> dict[str, Any]:
+    """Decide, without mutating anything, which registrations may be retired.
+
+    Every row lands in exactly one of ``candidates`` (eligible) or
+    ``preserved`` (with the reason code that stopped it), so the output
+    explains the whole registry rather than only the interesting part.
+    """
+    all_rows = [dict(row) for row in rows]
+    protected_ids = {str(w) for w in (protected_worker_identities or []) if w}
+    protected_sessions = {str(s) for s in (protected_session_ids or []) if s}
+    protected_pid_set = {_canon(p) for p in (protected_pids or []) if p is not None}
+
+    snapshots: dict[int, dict[str, Any]] = {}
+    pid_alive_by_index: dict[int, bool | None] = {}
+    start_time_by_index: dict[int, str | None] = {}
+    for index, row in enumerate(all_rows):
+        pid_alive = _probe_pid(row.get("pid"), pid_alive_probe)
+        pid_alive_by_index[index] = pid_alive
+        start_time_by_index[index] = _probe_start_time(
+            row.get("pid"), start_time_probe
+        )
+        snapshots[index] = fleet.build_worker_snapshot_row(
+            row,
+            now=now,
+            pid_alive_probe=(lambda _pid, _value=pid_alive: _value),
+            canonical_repository=canonical_repository,
+        )
+
+    # Identity evidence owned by a worker that is live, or whose liveness could
+    # not be established, is ambiguous: anything sharing it is preserved.
+    ambiguous_keys: set[tuple[str, str]] = set()
+    identity_counts: dict[str, int] = {}
+    for index, row in enumerate(all_rows):
+        identity = _canon(row.get("worker_identity"))
+        if identity:
+            identity_counts[identity] = identity_counts.get(identity, 0) + 1
+        snapshot_row = snapshots[index]
+        liveness = snapshot_row.get("liveness") or {}
+        unresolved = (
+            pid_alive_by_index[index] is None
+            or liveness.get("heartbeat_fresh") is None
+        )
+        if snapshot_row.get("live") or (
+            str(row.get("status") or "") == mwi.STATUS_ACTIVE and unresolved
+        ):
+            ambiguous_keys.update(_conflict_keys(row, snapshot_row))
+
+    # Two active registrations claiming one (client_instance_id, namespace)
+    # violate the #978 uniqueness invariant — but only when one of them might
+    # still be running. Several *dead* rows accumulating on one slot across
+    # restarts is ordinary history and every one of them is safely retirable;
+    # a slot shared with a live or unprobeable worker is genuinely ambiguous,
+    # because which row that process belongs to cannot be settled from the
+    # registry alone.
+    #
+    # The key is deliberately the (instance, namespace) pair, not the instance
+    # alone: one legitimate cohort is exactly one instance spread across
+    # distinct namespaces, so keying on the instance would make every cohort
+    # look self-conflicting and preserve the whole fleet forever.
+    instance_members: dict[tuple[str, str], list[int]] = {}
+    for index, row in enumerate(all_rows):
+        if str(row.get("status") or "") != mwi.STATUS_ACTIVE:
+            continue
+        key = _instance_key(row)
+        if key is None:
+            continue
+        instance_members.setdefault(key, []).append(index)
+    instance_conflicts: dict[tuple[str, str], bool] = {}
+    for key, members in instance_members.items():
+        if len(members) < 2:
+            continue
+        contested = any(
+            snapshots[i].get("live") or pid_alive_by_index[i] is None
+            for i in members
+        )
+        if contested:
+            instance_conflicts[key] = True
+
+    candidates: list[dict[str, Any]] = []
+    candidate_rows: list[Mapping[str, Any]] = []
+    preserved: list[dict[str, Any]] = []
+
+    for index, row in enumerate(all_rows):
+        snapshot_row = snapshots[index]
+        pid_alive = pid_alive_by_index[index]
+        liveness = snapshot_row.get("liveness") or {}
+        evidence = _evidence(row, snapshot_row, pid_alive)
+        blocked: tuple[str, str] | None = None
+
+        identity = _canon(row.get("worker_identity"))
+        missing = _missing_identity_fields(row)
+        shared = sorted(
+            f"{name}={value}"
+            for name, value in _conflict_keys(row, snapshot_row)
+            if (name, value) in ambiguous_keys
+        )
+        binding = (row.get("repository_binding") or "").strip()
+        protected_hits: list[str] = []
+        if identity and identity in protected_ids:
+            protected_hits.append(f"worker_identity={identity}")
+        if _canon(row.get("session_id")) in protected_sessions:
+            protected_hits.append(f"session_id={_canon(row.get('session_id'))}")
+        if _canon(row.get("pid")) in protected_pid_set:
+            protected_hits.append(f"pid={_canon(row.get('pid'))}")
+
+        if identity and identity_counts.get(identity, 0) > 1:
+            blocked = (
+                REASON_CONFLICTING_IDENTITY,
+                f"worker identity {identity!r} appears on more than one registry row",
+            )
+        elif str(row.get("status") or "") != mwi.STATUS_ACTIVE:
+            blocked = (
+                REASON_ALREADY_TERMINAL,
+                f"registration status is {row.get('status')!r}; nothing to retire",
+            )
+        elif missing:
+            blocked = (
+                REASON_INCOMPLETE_IDENTITY,
+                "registry row is missing field(s) the retirement conjunction "
+                f"reads: {missing}",
+            )
+        elif mwi._parse_ts(row.get("last_heartbeat_at")) is None:
+            blocked = (
+                REASON_UNPARSABLE_HEARTBEAT,
+                "last_heartbeat_at is not a parsable UTC stamp; liveness is unknown",
+            )
+        elif snapshot_row.get("live"):
+            blocked = (REASON_WORKER_LIVE, "worker is live and must not be retired")
+        elif pid_alive is None:
+            blocked = (
+                REASON_PID_UNKNOWN,
+                f"pid {row.get('pid')!r} could not be probed; liveness is unproven",
+            )
+        elif pid_alive and _reuse_detected(
+            row, start_time_by_index[index], current_host_id, current_boot_id
+        ):
+            # Reported before the generic live-pid branch so the operator sees
+            # *why* the number is occupied: an unrelated process inherited it,
+            # which means this registration's own liveness is unproven rather
+            # than positively established.
+            blocked = (
+                REASON_PID_REUSED,
+                f"pid {row.get('pid')!r} is alive but started at "
+                f"{start_time_by_index[index]!r}, not the registered "
+                f"{row.get('process_start_time')!r}; the number was reused",
+            )
+        elif pid_alive:
+            blocked = (
+                REASON_PID_ALIVE,
+                f"recorded pid {row.get('pid')!r} is still running",
+            )
+        elif liveness.get("heartbeat_fresh") is not False:
+            blocked = (
+                REASON_HEARTBEAT_FRESH,
+                "heartbeat has not expired under the canonical TTL policy",
+            )
+        elif snapshot_row.get("ownership_state") != "stale":
+            blocked = (
+                REASON_AMBIGUOUS_OWNERSHIP,
+                "ownership_state is "
+                f"{snapshot_row.get('ownership_state')!r}, not 'stale'",
+            )
+        elif not binding or snapshot_row.get("foreign_repository"):
+            blocked = (
+                REASON_FOREIGN_REPOSITORY,
+                "repository binding is absent or does not match the canonical "
+                "repository; retirement scope is ambiguous",
+            )
+        elif shared:
+            blocked = (
+                REASON_CONFLICTING_IDENTITY,
+                "identity evidence is shared with a live or unprobeable worker: "
+                f"{shared}",
+            )
+        elif protected_hits:
+            blocked = (
+                REASON_PROTECTED_OWNER,
+                "worker still owns active workflow state requiring separate "
+                f"reconciliation: {protected_hits}",
+            )
+        elif instance_conflicts.get(_instance_key(row)):
+            blocked = (
+                REASON_INSTANCE_CONFLICT,
+                "another active registration claims client_instance_id "
+                f"{row.get('client_instance_id')!r} in namespace "
+                f"{row.get('namespace')!r}; instance ownership is ambiguous",
+            )
+        else:
+            # Affirmative identity + fencing proof runs last: everything above
+            # establishes the row is *inert*, and this establishes it is
+            # unambiguously *this* worker (#980 review 657 B2).
+            blocked = assess_retirement_identity_proof(
+                row,
+                snapshot_row,
+                current_host_id=current_host_id,
+                current_boot_id=current_boot_id,
+                live_start_time=start_time_by_index[index],
+                pid_alive=pid_alive,
+            )
+
+        if blocked is not None:
+            preserved.append(
+                {
+                    "worker_identity": row.get("worker_identity"),
+                    "reason_code": blocked[0],
+                    "detail": blocked[1],
+                    "evidence": evidence,
+                }
+            )
+            continue
+
+        candidates.append(
+            {
+                "worker_identity": row.get("worker_identity"),
+                "reason_code": REASON_ELIGIBLE,
+                "detail": (
+                    "dead pid, expired heartbeat, stale ownership, unambiguous "
+                    "identity, canonical repository binding, no active workflow "
+                    "ownership"
+                ),
+                "evidence": evidence,
+            }
+        )
+        candidate_rows.append(row)
+
+    counts: dict[str, int] = {}
+    for entry in preserved:
+        counts[entry["reason_code"]] = counts.get(entry["reason_code"], 0) + 1
+
+    return {
+        "success": True,
+        "read_only": True,
+        "mutation_performed": False,
+        "outcome": OUTCOME_PLANNED,
+        "registry_fingerprint": registry_fingerprint(all_rows),
+        "candidate_fingerprint": candidate_fingerprint(candidate_rows),
+        "assessed_count": len(all_rows),
+        "candidate_count": len(candidates),
+        "preserved_count": len(preserved),
+        "candidates": candidates,
+        "candidate_worker_identities": [c["worker_identity"] for c in candidates],
+        "preserved": preserved,
+        "preserved_reason_counts": counts,
+        "protected_inputs": {
+            "worker_identities": sorted(protected_ids),
+            "session_ids": sorted(protected_sessions),
+            "pids": sorted(protected_pid_set),
+        },
+        "canonical_repository": canonical_repository,
+        "fencing_context": {
+            "current_host_id": current_host_id,
+            "current_boot_id": current_boot_id,
+            "start_time_probe_available": start_time_probe is not None,
+        },
+    }
+
+
+def summarize_plan(plan: Mapping[str, Any]) -> dict[str, Any]:
+    """Compact, log-safe view of a plan or apply result."""
+    return {
+        "outcome": plan.get("outcome"),
+        "registry_fingerprint": plan.get("registry_fingerprint"),
+        "candidate_fingerprint": plan.get("candidate_fingerprint"),
+        "assessed_count": plan.get("assessed_count"),
+        "candidate_count": plan.get("candidate_count"),
+        "retired_count": plan.get("retired_count"),
+        "preserved_count": plan.get("preserved_count"),
+        "mutation_performed": plan.get("mutation_performed"),
+    }

mcp_fleet_snapshot.py

@@ -0,0 +1,869 @@
+"""Instance-level fleet identity and health snapshots (#978).
+
+#948 established per-worker ownership; #975 made heartbeats keep those rows
+live. Neither surface could enumerate the fleet at *instance* granularity:
+which application launch owns which five namespace workers, whether two
+Codex launches are distinct, or whether a live collision is real rather than
+a shared client type.
+
+This module is pure. Callers supply registry rows (and optional enrichments);
+nothing here opens SQLite, scans process tables, or mutates state. Production
+evidence for the fleet gate is the snapshot returned by the sanctioned
+controller/reconciler tool that wraps this assessor.
+
+Identity hierarchy (highest → lowest):
+
+* ``client_type`` — application family (``codex``, ``claude_code``, …)
+* ``client_instance_id`` — one running application launch (trusted launcher)
+* ``fleet_run_id`` — operator-approved enrollment / canary cohort
+* ``worker_id`` / ``worker_identity`` — one namespace worker process
+* ``namespace`` — author | reviewer | merger | controller | reconciler
+
+Multiple simultaneous instances of the same ``client_type`` are first-class.
+Sharing only a profile or client type is never a duplicate.
+"""
+
+from __future__ import annotations
+
+import hashlib
+import secrets
+from collections import defaultdict
+from datetime import datetime, timezone
+from typing import Any, Callable, Iterable, Mapping
+
+import mcp_worker_identity as mwi
+
+# --- Classification labels ------------------------------------------------
+
+CLASS_EXPECTED = "expected_enrolled"
+CLASS_MISSING = "missing_expected"
+CLASS_UNMANIFESTED = "unmanifested"
+CLASS_DUPLICATE_NAMESPACE = "duplicate_namespace_worker"
+CLASS_INSTANCE_ID_COLLISION = "instance_id_collision"
+CLASS_WORKER_ID_COLLISION = "worker_identity_collision"
+CLASS_SESSION_COLLISION = "session_identity_collision"
+CLASS_GENERATION_COLLISION = "generation_identity_collision"
+CLASS_PROCESS_COLLISION = "process_identity_collision"
+CLASS_PID_COLLISION = "pid_collision"
+CLASS_OWNERSHIP_COLLISION = "ownership_fencing_collision"
+CLASS_ORPHANED = "orphaned_unowned"
+CLASS_UNKNOWN_CLIENT = "unknown_client"
+CLASS_FOREIGN_REPOSITORY = "foreign_repository"
+CLASS_OLD_REVISION = "old_revision"
+CLASS_STALE_WORKER = "stale_orphaned_worker"
+CLASS_LEGACY_INCOMPLETE = "legacy_incomplete_identity"
+CLASS_HISTORICAL = "historical_dead"
+CLASS_HEALTHY = "healthy"
+
+#: Active blockers that make the live fleet unsafe for mutation-gated work.
+ACTIVE_BLOCKER_CLASSES = frozenset(
+    {
+        CLASS_MISSING,
+        CLASS_UNMANIFESTED,
+        CLASS_DUPLICATE_NAMESPACE,
+        CLASS_INSTANCE_ID_COLLISION,
+        CLASS_WORKER_ID_COLLISION,
+        CLASS_SESSION_COLLISION,
+        CLASS_GENERATION_COLLISION,
+        CLASS_PROCESS_COLLISION,
+        CLASS_PID_COLLISION,
+        CLASS_OWNERSHIP_COLLISION,
+        CLASS_ORPHANED,
+        CLASS_UNKNOWN_CLIENT,
+        CLASS_FOREIGN_REPOSITORY,
+        CLASS_OLD_REVISION,
+        CLASS_STALE_WORKER,
+        CLASS_LEGACY_INCOMPLETE,
+    }
+)
+
+SANCTIONED_NAMESPACES = frozenset(
+    {"author", "reviewer", "merger", "controller", "reconciler"}
+)
+
+INSTANCE_ID_PROVENANCE_TRUSTED = "trusted_launcher"
+INSTANCE_ID_PROVENANCE_LEGACY = "legacy_incomplete"
+INSTANCE_ID_PROVENANCE_MISSING = "missing"
+
+CLIENT_INSTANCE_ENV = "GITEA_MCP_CLIENT_INSTANCE"
+FLEET_RUN_ENV = "GITEA_MCP_FLEET_RUN_ID"
+PROCESS_IDENTITY_ENV = "GITEA_MCP_PROCESS_IDENTITY"
+INSTANCE_PROVENANCE_ENV = "GITEA_MCP_INSTANCE_PROVENANCE"
+
+_LEGACY_INSTANCE_PREFIXES = ("pid-", "proc-", "legacy-")
+#: Trusted launcher-issued IDs use the reserved ``inst-`` prefix
+#: (see :func:`generate_client_instance_id`). Ordinary user-supplied strings
+#: without that prefix never count as trusted attribution.
+_TRUSTED_INSTANCE_PREFIX = "inst-"
+
+
+def _utc_now() -> datetime:
+    return datetime.now(timezone.utc)
+
+
+def _ts(value: datetime) -> str:
+    return value.astimezone(timezone.utc).strftime("%Y-%m-%dT%H:%M:%SZ")
+
+
+def generate_client_instance_id(
+    client_type: str | None,
+    *,
+    launch_nonce: str | None = None,
+    now: datetime | None = None,
+) -> str:
+    """Mint a distinct instance ID for one application launch (#978).
+
+    The trusted launcher (or host that starts all five namespace workers)
+    generates this once per launch and injects it as ``GITEA_MCP_CLIENT_INSTANCE``
+    into every worker environment. Workers never invent their own instance ID
+    from PID proximity or timestamps.
+    """
+    client = mwi.normalize_client_name(client_type)
+    stamp = (now or _utc_now()).astimezone(timezone.utc).strftime(
+        mwi.IDENTITY_TIMESTAMP_FORMAT
+    )
+    nonce = launch_nonce if launch_nonce is not None else secrets.token_hex(16)
+    digest = hashlib.sha256(
+        f"{client}\x1f{stamp}\x1f{nonce}".encode("utf-8")
+    ).hexdigest()[:12]
+    return f"inst-{client}-{stamp}-{digest}"
+
+
+def assess_instance_identity(
+    raw_instance_id: str | None,
+    *,
+    source: str | None = None,
+) -> dict[str, Any]:
+    """Classify whether a client_instance_id is trusted enough for mutation.
+
+    Trusted instance IDs are non-empty, match the launcher-minted ``inst-…``
+    format from :func:`generate_client_instance_id`, and are not pre-#978
+    PID/proc/legacy placeholders. Ordinary user-supplied or malformed values
+    fail closed as untrusted so they cannot spoof multi-instance attribution.
+    Incomplete identities remain visible for diagnosis but cannot authorize
+    unsafe mutation.
+    """
+    text = (raw_instance_id or "").strip()
+    if not text:
+        return {
+            "client_instance_id": None,
+            "complete": False,
+            "trusted": False,
+            "provenance": INSTANCE_ID_PROVENANCE_MISSING,
+            "reasons": [
+                "client_instance_id is missing; the trusted launcher must set "
+                f"{CLIENT_INSTANCE_ENV} once per application launch"
+            ],
+        }
+    lowered = text.lower()
+    if lowered.startswith(_LEGACY_INSTANCE_PREFIXES) or source == "pid_fallback":
+        return {
+            "client_instance_id": text,
+            "complete": False,
+            "trusted": False,
+            "provenance": INSTANCE_ID_PROVENANCE_LEGACY,
+            "reasons": [
+                f"client_instance_id {text!r} is a legacy PID/process fallback, "
+                "not a trusted launcher-issued instance identity"
+            ],
+        }
+    if not text.startswith(_TRUSTED_INSTANCE_PREFIX) or len(text) <= len(
+        _TRUSTED_INSTANCE_PREFIX
+    ):
+        return {
+            "client_instance_id": text,
+            "complete": False,
+            "trusted": False,
+            "provenance": INSTANCE_ID_PROVENANCE_LEGACY,
+            "reasons": [
+                f"client_instance_id {text!r} is malformed or user-supplied and "
+                f"does not use the trusted launcher prefix "
+                f"{_TRUSTED_INSTANCE_PREFIX!r}; refusing trusted attribution"
+            ],
+        }
+    return {
+        "client_instance_id": text,
+        "complete": True,
+        "trusted": True,
+        "provenance": INSTANCE_ID_PROVENANCE_TRUSTED,
+        "reasons": [],
+    }
+
+
+def resolve_client_instance_from_env(
+    env: Mapping[str, str] | None = None,
+    *,
+    pid: int | None = None,
+) -> dict[str, Any]:
+    """Resolve instance identity from launcher env without inventing one.
+
+    When the trusted key is absent, return incomplete evidence rather than a
+    silent ``pid-<n>`` identity. Callers that still need a non-empty registry
+    key may choose a legacy placeholder deliberately; they must not treat it as
+    trusted.
+    """
+    source = dict(env or {})
+    raw = (source.get(CLIENT_INSTANCE_ENV) or "").strip()
+    assessment = assess_instance_identity(raw or None)
+    assessment["fleet_run_id"] = (source.get(FLEET_RUN_ENV) or "").strip() or None
+    assessment["process_identity"] = (
+        (source.get(PROCESS_IDENTITY_ENV) or "").strip()
+        or (f"pid-{pid}" if pid is not None else None)
+    )
+    return assessment
+
+
+def _public_worker(record: Mapping[str, Any]) -> dict[str, Any]:
+    return {
+        "worker_id": record.get("worker_identity") or record.get("worker_id"),
+        "worker_identity": record.get("worker_identity") or record.get("worker_id"),
+        "client_type": mwi.normalize_client_name(
+            record.get("client_name") or record.get("client_type")
+        ),
+        "client_instance_id": record.get("client_instance_id"),
+        "fleet_run_id": record.get("fleet_run_id"),
+        "namespace": record.get("namespace"),
+        "profile": record.get("profile"),
+        "declared_role": record.get("role") or record.get("declared_role"),
+        "authenticated_account": record.get("authenticated_account"),
+        "session_id": record.get("session_id"),
+        "generation_id": record.get("generation_id"),
+        "process_identity": record.get("process_identity")
+        or (
+            f"pid-{record['pid']}"
+            if record.get("pid") is not None
+            else None
+        ),
+        "pid": record.get("pid"),
+        "repository_binding": record.get("repository_binding"),
+        "remote": record.get("remote"),
+        "startup_revision": record.get("startup_revision"),
+        "loaded_revision": record.get("loaded_revision"),
+        "parity_revision": record.get("parity_revision"),
+        "live_revision": record.get("live_revision"),
+        "runtime_provenance": record.get("runtime_provenance")
+        or record.get("transport"),
+        "transport": record.get("transport"),
+        "started_at": record.get("started_at"),
+        "last_heartbeat_at": record.get("last_heartbeat_at"),
+        "heartbeat_ttl_seconds": record.get("heartbeat_ttl_seconds"),
+        "fencing_epoch": record.get("fencing_epoch"),
+        "status": record.get("status"),
+        "instance_id_provenance": record.get("instance_id_provenance"),
+    }
+
+
+def _liveness(
+    record: Mapping[str, Any],
+    *,
+    now: datetime | None,
+    pid_alive_probe: Callable[[int | None], bool | None] | None,
+) -> dict[str, Any]:
+    pid_alive = None
+    if pid_alive_probe is not None and record.get("pid") is not None:
+        try:
+            pid_alive = pid_alive_probe(record.get("pid"))
+        except Exception:
+            pid_alive = None
+    return mwi.WorkerRegistry.is_live(record, now=now, pid_alive=pid_alive)
+
+
+def _consistency_token(rows: Iterable[Mapping[str, Any]], snapshot_at: str) -> str:
+    material = [snapshot_at]
+    for row in sorted(
+        rows,
+        key=lambda r: (
+            str(r.get("worker_identity") or ""),
+            str(r.get("last_heartbeat_at") or ""),
+            str(r.get("fencing_epoch") or ""),
+        ),
+    ):
+        material.append(
+            "|".join(
+                [
+                    str(row.get("worker_identity") or ""),
+                    str(row.get("client_instance_id") or ""),
+                    str(row.get("status") or ""),
+                    str(row.get("last_heartbeat_at") or ""),
+                    str(row.get("fencing_epoch") or ""),
+                    str(row.get("generation_id") or ""),
+                ]
+            )
+        )
+    digest = hashlib.sha256("\n".join(material).encode("utf-8")).hexdigest()[:16]
+    return f"fleetrev-{digest}"
+
+
+def build_worker_snapshot_row(
+    record: Mapping[str, Any],
+    *,
+    now: datetime | None = None,
+    pid_alive_probe: Callable[[int | None], bool | None] | None = None,
+    canonical_repository: str | None = None,
+    expected_live_revision: str | None = None,
+    heartbeat_supervised: bool | None = None,
+) -> dict[str, Any]:
+    """One point-in-time worker row for the fleet snapshot."""
+    stamp = now or _utc_now()
+    base = _public_worker(record)
+    identity = assess_instance_identity(
+        base.get("client_instance_id"),
+        source=record.get("instance_id_source"),
+    )
+    liveness = _liveness(record, now=stamp, pid_alive_probe=pid_alive_probe)
+    is_historical = str(record.get("status") or "") != mwi.STATUS_ACTIVE
+    live = bool(liveness.get("live")) and not is_historical
+
+    repo = (base.get("repository_binding") or "").strip() or None
+    foreign_repo = bool(
+        canonical_repository
+        and repo
+        and repo.rstrip("/") != str(canonical_repository).rstrip("/")
+    )
+    old_revision = False
+    if expected_live_revision:
+        for key in ("startup_revision", "loaded_revision", "parity_revision", "live_revision"):
+            rev = (base.get(key) or "").strip()
+            if rev and rev != expected_live_revision:
+                old_revision = True
+                break
+
+    ownership_state = "historical" if is_historical else (
+        "live" if live else "stale"
+    )
+    if live and not identity["trusted"]:
+        ownership_state = "live_untrusted_identity"
+    if live and not base.get("session_id"):
+        ownership_state = "orphaned"
+
+    mutation_safe = bool(
+        live
+        and identity["trusted"]
+        and not foreign_repo
+        and not old_revision
+        and ownership_state == "live"
+        and base.get("client_type") != mwi.UNKNOWN_CLIENT
+    )
+
+    restart_required = bool(
+        old_revision
+        or (live and not liveness.get("heartbeat_fresh", True))
+    )
+
+    return {
+        **base,
+        "instance_identity": identity,
+        "client_instance_id": identity["client_instance_id"] or base.get("client_instance_id"),
+        "instance_id_provenance": identity["provenance"],
+        "instance_identity_trusted": identity["trusted"],
+        "live": live,
+        "historical": is_historical,
+        "liveness": liveness,
+        "heartbeat": {
+            "registered": bool(base.get("last_heartbeat_at")),
+            "supervised": heartbeat_supervised,
+            "age_seconds": liveness.get("heartbeat_age_seconds"),
+            "ttl_seconds": liveness.get("heartbeat_ttl_seconds"),
+            "fresh": liveness.get("heartbeat_fresh"),
+            "last_heartbeat_at": base.get("last_heartbeat_at"),
+        },
+        "fencing": {
+            "fencing_epoch": base.get("fencing_epoch"),
+            "generation_id": base.get("generation_id"),
+        },
+        "ownership_state": ownership_state,
+        "foreign_repository": foreign_repo,
+        "old_revision": old_revision,
+        "stale": not live and not is_historical,
+        "restart_required": restart_required,
+        "mutation_safe": mutation_safe,
+        "conflicting_live_sessions": [],
+    }
+
+
+def _collision_groups(
+    live_rows: list[dict[str, Any]],
+    key_fn,
+) -> dict[str, list[dict[str, Any]]]:
+    groups: dict[str, list[dict[str, Any]]] = defaultdict(list)
+    for row in live_rows:
+        key = key_fn(row)
+        if key is None or key == "" or key == "None":
+            continue
+        groups[str(key)].append(row)
+    return {k: v for k, v in groups.items() if len(v) > 1}
+
+
+def snapshot_instance_fleet(
+    records: Iterable[Mapping[str, Any]],
+    *,
+    expected_manifest: list[Mapping[str, Any]] | None = None,
+    now: datetime | None = None,
+    pid_alive_probe: Callable[[int | None], bool | None] | None = None,
+    canonical_repository: str | None = None,
+    expected_live_revision: str | None = None,
+    registry_revision: str | None = None,
+    known_client_types: Iterable[str] | None = None,
+) -> dict[str, Any]:
+    """Authoritative point-in-time fleet snapshot with classification (#978).
+
+    Historical dead rows are reported separately and never automatically make
+    the live fleet unsafe.
+    """
+    stamp = now or _utc_now()
+    snapshot_at = _ts(stamp)
+    known = {
+        mwi.normalize_client_name(c)
+        for c in (known_client_types or mwi.CLIENT_ALIASES.values())
+    }
+    known.discard(mwi.UNKNOWN_CLIENT)
+
+    all_rows: list[dict[str, Any]] = []
+    for record in records:
+        all_rows.append(
+            build_worker_snapshot_row(
+                record,
+                now=stamp,
+                pid_alive_probe=pid_alive_probe,
+                canonical_repository=canonical_repository,
+                expected_live_revision=expected_live_revision,
+            )
+        )
+
+    live_rows = [r for r in all_rows if r["live"]]
+    historical_rows = [r for r in all_rows if r["historical"]]
+    stale_rows = [r for r in all_rows if r["stale"]]
+
+    # --- identity collisions among live workers ---
+    findings: list[dict[str, Any]] = []
+
+    def _finding(
+        classification: str,
+        *,
+        severity: str,
+        workers: list[dict[str, Any]] | None = None,
+        instance_ids: list[str] | None = None,
+        detail: str,
+        active_blocker: bool,
+    ) -> None:
+        findings.append(
+            {
+                "classification": classification,
+                "severity": severity,
+                "active_blocker": active_blocker,
+                "detail": detail,
+                "client_instance_ids": instance_ids or sorted(
+                    {
+                        str(w.get("client_instance_id"))
+                        for w in (workers or [])
+                        if w.get("client_instance_id")
+                    }
+                ),
+                "worker_identities": [
+                    w.get("worker_identity") for w in (workers or [])
+                ],
+            }
+        )
+
+    # Duplicate worker identity (should not happen with PK, still detect)
+    for wid, group in _collision_groups(
+        live_rows, lambda r: r.get("worker_identity")
+    ).items():
+        _finding(
+            CLASS_WORKER_ID_COLLISION,
+            severity="blocker",
+            workers=group,
+            detail=f"worker identity {wid!r} is claimed by {len(group)} live workers",
+            active_blocker=True,
+        )
+
+    # Reused session identity across live workers
+    for sid, group in _collision_groups(live_rows, lambda r: r.get("session_id")).items():
+        # Same session may appear once; collision only when multiple workers share it
+        # across different worker identities (always true for group size > 1).
+        _finding(
+            CLASS_SESSION_COLLISION,
+            severity="blocker",
+            workers=group,
+            detail=f"session identity {sid!r} is reused by {len(group)} live workers",
+            active_blocker=True,
+        )
+
+    # Generation claimed by multiple live sessions/workers is a conflict when
+    # the workers are not the five sanctioned namespaces of one instance.
+    for gen, group in _collision_groups(
+        live_rows, lambda r: r.get("generation_id")
+    ).items():
+        namespaces = {g.get("namespace") for g in group if g.get("namespace")}
+        instance_ids = {g.get("client_instance_id") for g in group}
+        # Multiple workers under one generation is only valid if they share one
+        # instance and distinct namespaces. Same generation + same namespace = bad.
+        by_ns: dict[str, list] = defaultdict(list)
+        for g in group:
+            by_ns[str(g.get("namespace") or "")].append(g)
+        ns_dups = {ns: rows for ns, rows in by_ns.items() if ns and len(rows) > 1}
+        if ns_dups or len(instance_ids) > 1:
+            _finding(
+                CLASS_GENERATION_COLLISION,
+                severity="blocker",
+                workers=group,
+                detail=(
+                    f"generation {gen!r} is contested across namespaces/instances "
+                    f"(namespaces={sorted(namespaces)}, "
+                    f"instances={sorted(str(i) for i in instance_ids if i)})"
+                ),
+                active_blocker=True,
+            )
+
+    # Process identity / PID collisions across distinct workers
+    for proc, group in _collision_groups(
+        live_rows, lambda r: r.get("process_identity")
+    ).items():
+        if len({r.get("worker_identity") for r in group}) > 1:
+            _finding(
+                CLASS_PROCESS_COLLISION,
+                severity="blocker",
+                workers=group,
+                detail=f"process identity {proc!r} is shared by distinct live workers",
+                active_blocker=True,
+            )
+
+    for pid, group in _collision_groups(live_rows, lambda r: r.get("pid")).items():
+        if len({r.get("worker_identity") for r in group}) > 1:
+            _finding(
+                CLASS_PID_COLLISION,
+                severity="blocker",
+                workers=group,
+                detail=f"PID {pid} is shared by distinct live workers",
+                active_blocker=True,
+            )
+
+    # Fencing/ownership: same fencing epoch on different workers of different instances
+    for epoch, group in _collision_groups(
+        live_rows,
+        lambda r: (
+            f"{r.get('generation_id')}:{r.get('fencing_epoch')}"
+            if r.get("generation_id") is not None and r.get("fencing_epoch") is not None
+            else None
+        ),
+    ).items():
+        if len({r.get("client_instance_id") for r in group}) > 1:
+            _finding(
+                CLASS_OWNERSHIP_COLLISION,
+                severity="blocker",
+                workers=group,
+                detail=(
+                    f"fencing token {epoch!r} spans more than one client_instance_id"
+                ),
+                active_blocker=True,
+            )
+
+    # Per-instance grouping
+    by_instance: dict[str, list[dict[str, Any]]] = defaultdict(list)
+    unkeyed_live: list[dict[str, Any]] = []
+    for row in live_rows:
+        iid = row.get("client_instance_id")
+        if not iid:
+            unkeyed_live.append(row)
+            continue
+        by_instance[str(iid)].append(row)
+
+    instances: list[dict[str, Any]] = []
+    for iid, workers in sorted(by_instance.items()):
+        client_types = sorted({w.get("client_type") for w in workers if w.get("client_type")})
+        trusted = all(w.get("instance_identity_trusted") for w in workers)
+        namespaces = [w.get("namespace") for w in workers]
+        ns_counts: dict[str, int] = defaultdict(int)
+        for ns in namespaces:
+            if ns:
+                ns_counts[str(ns)] += 1
+        dup_ns = sorted(ns for ns, n in ns_counts.items() if n > 1)
+        if dup_ns:
+            _finding(
+                CLASS_DUPLICATE_NAMESPACE,
+                severity="blocker",
+                workers=[w for w in workers if w.get("namespace") in dup_ns],
+                instance_ids=[iid],
+                detail=(
+                    f"instance {iid!r} has more than one live worker for "
+                    f"namespace(s) {dup_ns}"
+                ),
+                active_blocker=True,
+            )
+
+        # Live reuse of one instance ID with incompatible client types
+        if len(client_types) > 1:
+            _finding(
+                CLASS_INSTANCE_ID_COLLISION,
+                severity="blocker",
+                workers=workers,
+                instance_ids=[iid],
+                detail=(
+                    f"client_instance_id {iid!r} is live under multiple client "
+                    f"types {client_types}"
+                ),
+                active_blocker=True,
+            )
+
+        if not trusted:
+            _finding(
+                CLASS_LEGACY_INCOMPLETE,
+                severity="blocker",
+                workers=workers,
+                instance_ids=[iid],
+                detail=(
+                    f"instance {iid!r} lacks trusted launcher-issued instance "
+                    "identity; diagnostic reads remain available"
+                ),
+                active_blocker=True,
+            )
+
+        unknown = [w for w in workers if w.get("client_type") == mwi.UNKNOWN_CLIENT]
+        if unknown:
+            _finding(
+                CLASS_UNKNOWN_CLIENT,
+                severity="blocker",
+                workers=unknown,
+                instance_ids=[iid],
+                detail=f"instance {iid!r} has worker(s) with unknown client_type",
+                active_blocker=True,
+            )
+
+        foreign = [w for w in workers if w.get("foreign_repository")]
+        if foreign:
+            _finding(
+                CLASS_FOREIGN_REPOSITORY,
+                severity="blocker",
+                workers=foreign,
+                instance_ids=[iid],
+                detail=f"instance {iid!r} has foreign-repository workers",
+                active_blocker=True,
+            )
+
+        old = [w for w in workers if w.get("old_revision")]
+        if old:
+            _finding(
+                CLASS_OLD_REVISION,
+                severity="blocker",
+                workers=old,
+                instance_ids=[iid],
+                detail=f"instance {iid!r} has old-revision workers",
+                active_blocker=True,
+            )
+
+        orphans = [w for w in workers if w.get("ownership_state") == "orphaned"]
+        if orphans:
+            _finding(
+                CLASS_ORPHANED,
+                severity="blocker",
+                workers=orphans,
+                instance_ids=[iid],
+                detail=f"instance {iid!r} has orphaned/unowned workers",
+                active_blocker=True,
+            )
+
+        instances.append(
+            {
+                "client_instance_id": iid,
+                "client_types": client_types,
+                "client_type": client_types[0] if len(client_types) == 1 else None,
+                "fleet_run_ids": sorted(
+                    {w.get("fleet_run_id") for w in workers if w.get("fleet_run_id")}
+                ),
+                "worker_count": len(workers),
+                "namespaces": sorted({n for n in namespaces if n}),
+                "namespace_counts": dict(ns_counts),
+                "duplicate_namespaces": dup_ns,
+                "trusted_instance_identity": trusted,
+                "workers": workers,
+                "mutation_safe": all(w.get("mutation_safe") for w in workers)
+                and not dup_ns
+                and trusted,
+            }
+        )
+
+    for row in unkeyed_live:
+        _finding(
+            CLASS_LEGACY_INCOMPLETE,
+            severity="blocker",
+            workers=[row],
+            detail="live worker has no client_instance_id",
+            active_blocker=True,
+        )
+        if row.get("client_type") == mwi.UNKNOWN_CLIENT:
+            _finding(
+                CLASS_UNKNOWN_CLIENT,
+                severity="blocker",
+                workers=[row],
+                detail="live worker has unknown client_type and no instance id",
+                active_blocker=True,
+            )
+
+    for row in stale_rows:
+        _finding(
+            CLASS_STALE_WORKER,
+            severity="warning",
+            workers=[row],
+            detail=(
+                f"worker {row.get('worker_identity')!r} is active in the registry "
+                "but not live (stale heartbeat or dead pid)"
+            ),
+            active_blocker=True,
+        )
+
+    for row in historical_rows:
+        _finding(
+            CLASS_HISTORICAL,
+            severity="info",
+            workers=[row],
+            detail=(
+                f"historical registration {row.get('worker_identity')!r} "
+                f"(status={row.get('status')!r}) is not an active blocker"
+            ),
+            active_blocker=False,
+        )
+
+    # Manifest comparison
+    expected = list(expected_manifest or [])
+    expected_ids = {
+        str(item.get("client_instance_id")).strip()
+        for item in expected
+        if (item.get("client_instance_id") or "").strip()
+    }
+    live_ids = set(by_instance.keys())
+    missing_ids = sorted(expected_ids - live_ids)
+    unmanifested_ids = sorted(live_ids - expected_ids) if expected_ids else []
+
+    for iid in missing_ids:
+        _finding(
+            CLASS_MISSING,
+            severity="blocker",
+            instance_ids=[iid],
+            detail=f"expected enrolled instance {iid!r} is missing from the live fleet",
+            active_blocker=True,
+        )
+    for iid in unmanifested_ids:
+        _finding(
+            CLASS_UNMANIFESTED,
+            severity="blocker",
+            instance_ids=[iid],
+            workers=by_instance.get(iid, []),
+            detail=(
+                f"live instance {iid!r} is not on the approved fleet manifest "
+                "(unmanifested)"
+            ),
+            active_blocker=True,
+        )
+
+    # Same client_type multi-instance is healthy when each has distinct instance IDs
+    by_type: dict[str, list[str]] = defaultdict(list)
+    for inst in instances:
+        for ct in inst.get("client_types") or []:
+            by_type[str(ct)].append(inst["client_instance_id"])
+    multi_instance_same_type = {
+        ct: ids for ct, ids in by_type.items() if len(ids) > 1
+    }
+
+    active_blockers = [f for f in findings if f.get("active_blocker")]
+    historical_only = [f for f in findings if f.get("classification") == CLASS_HISTORICAL]
+    live_safe = not active_blockers
+
+    consistency = registry_revision or _consistency_token(all_rows, snapshot_at)
+
+    return {
+        "success": True,
+        "read_only": True,
+        "snapshot_at": snapshot_at,
+        "consistency_token": consistency,
+        "registry_revision": consistency,
+        "live_worker_count": len(live_rows),
+        "historical_worker_count": len(historical_rows),
+        "stale_worker_count": len(stale_rows),
+        "instance_count": len(instances),
+        "workers": all_rows,
+        "live_workers": live_rows,
+        "historical_workers": historical_rows,
+        "stale_workers": stale_rows,
+        "instances": instances,
+        "multi_instance_same_client_type": multi_instance_same_type,
+        "same_client_type_not_duplicate": True,
+        "expected_manifest": [
+            {
+                "client_instance_id": item.get("client_instance_id"),
+                "client_type": item.get("client_type"),
+                "fleet_run_id": item.get("fleet_run_id"),
+                "namespaces": item.get("namespaces"),
+            }
+            for item in expected
+        ],
+        "missing_expected_instance_ids": missing_ids,
+        "unmanifested_instance_ids": unmanifested_ids,
+        "findings": findings,
+        "active_blockers": active_blockers,
+        "historical_findings": historical_only,
+        "live_fleet_safe": live_safe,
+        "mutation_safe": live_safe and all(
+            inst.get("mutation_safe") for inst in instances
+        )
+        if instances
+        else live_safe,
+        "classification_model": {
+            "exactly_one_process_per_profile": False,
+            "exactly_one_instance_per_client_type": False,
+            "multiple_instances_per_client_type": True,
+            "duplicate_requires": [
+                "live client_instance_id collision",
+                "duplicate namespace worker within one instance",
+                "reused worker/session/generation/process/pid/fencing identity",
+                "cross-instance ownership collision",
+                "unmanifested instance when a manifest is required",
+            ],
+        },
+        "reasons": [f["detail"] for f in active_blockers],
+    }
+
+
+def compare_snapshot_heartbeats(
+    earlier: Mapping[str, Any],
+    later: Mapping[str, Any],
+) -> dict[str, Any]:
+    """Prove heartbeat continuity and stable ownership across two snapshots."""
+    earlier_live = {
+        w.get("worker_identity"): w for w in earlier.get("live_workers") or []
+    }
+    later_live = {
+        w.get("worker_identity"): w for w in later.get("live_workers") or []
+    }
+    shared = sorted(set(earlier_live) & set(later_live))
+    continuity: list[dict[str, Any]] = []
+    stable_ownership = True
+    for wid in shared:
+        a = earlier_live[wid]
+        b = later_live[wid]
+        same_instance = a.get("client_instance_id") == b.get("client_instance_id")
+        same_session = a.get("session_id") == b.get("session_id")
+        same_generation = a.get("generation_id") == b.get("generation_id")
+        hb_advanced_or_equal = True
+        if a.get("last_heartbeat_at") and b.get("last_heartbeat_at"):
+            hb_advanced_or_equal = b["last_heartbeat_at"] >= a["last_heartbeat_at"]
+        if not (same_instance and same_session and same_generation):
+            stable_ownership = False
+        continuity.append(
+            {
+                "worker_identity": wid,
+                "same_client_instance_id": same_instance,
+                "same_session_id": same_session,
+                "same_generation_id": same_generation,
+                "heartbeat_non_decreasing": hb_advanced_or_equal,
+                "earlier_heartbeat": a.get("last_heartbeat_at"),
+                "later_heartbeat": b.get("last_heartbeat_at"),
+            }
+        )
+    return {
+        "shared_live_workers": shared,
+        "continuity": continuity,
+        "stable_ownership": stable_ownership
+        and all(c["heartbeat_non_decreasing"] for c in continuity),
+        "dropped_workers": sorted(set(earlier_live) - set(later_live)),
+        "new_workers": sorted(set(later_live) - set(earlier_live)),
+    }

mcp_process_fencing.py

@@ -0,0 +1,205 @@
+"""Host, boot, and process-start fencing evidence for retirement safety (#980).
+
+Review 657 B2/B3 established that a local ``os.kill(pid, 0)`` probe is not, on
+its own, evidence that a *particular registered worker* is gone:
+
+* **Host ambiguity.** A registry row written on host A records pid 1234. Probing
+  pid 1234 on host B answers a question nobody asked. "Not running here" is not
+  "not running".
+* **PID reuse.** Pid 1234 may be alive and belong to an unrelated process that
+  the kernel handed the number to after the original exited. The naive probe
+  reads that as "the worker is live" (safe, over-preserving) — but the converse
+  matters too: evidence captured about pid 1234 at time T must not be honoured
+  at time T+n if the process behind that number changed in between.
+* **Boot boundaries.** Every pid from a previous boot is conclusively gone, and
+  pid numbers restart, so a recorded pid is only comparable to a live pid when
+  both belong to the same boot.
+
+This module supplies the three pieces of evidence that turn a bare pid into a
+statement about one specific process:
+
+``host_id``
+    Which machine the pid belongs to.
+``boot_id``
+    Which boot of that machine the pid belongs to. Pids are only comparable
+    within a single boot.
+``process_start_time``
+    Which *incarnation* of that pid number. Two processes on the same host and
+    boot sharing a pid number cannot share a start time, so comparing start
+    times defeats reuse.
+
+Every probe here is read-only, never raises, and returns ``None`` when the
+evidence cannot be established. ``None`` means *unknown*, and the retirement
+conjunction is required to treat unknown as "preserve", never as "safe".
+"""
+
+from __future__ import annotations
+
+import os
+import platform
+import subprocess
+
+# --- Host identity --------------------------------------------------------
+
+
+def current_host_id() -> str | None:
+    """Stable identifier for the machine this process runs on.
+
+    Deliberately the kernel node name rather than anything network-derived: it
+    does not change when an interface goes down or a VPN reassigns an address,
+    and retirement must not become unsafe because DNS moved.
+    """
+    try:
+        node = (platform.node() or "").strip()
+    except Exception:
+        return None
+    return node or None
+
+
+# --- Boot identity --------------------------------------------------------
+
+
+def _linux_boot_id() -> str | None:
+    try:
+        with open("/proc/sys/kernel/random/boot_id", encoding="utf-8") as handle:
+            value = handle.read().strip()
+    except Exception:
+        return None
+    return value or None
+
+
+def _darwin_boot_id() -> str | None:
+    """macOS boot identity, derived from ``kern.boottime``.
+
+    ``sysctl`` prints e.g. ``{ sec = 1785400000, usec = 123456 } Wed Jul 30 ...``.
+    Only the integer seconds are kept: the trailing human-readable date is
+    locale-dependent and would make the identifier unstable across environments
+    for the very same boot.
+    """
+    try:
+        completed = subprocess.run(
+            ["/usr/sbin/sysctl", "-n", "kern.boottime"],
+            capture_output=True,
+            text=True,
+            timeout=5,
+            check=False,
+        )
+    except Exception:
+        return None
+    if completed.returncode != 0:
+        return None
+    text = (completed.stdout or "").strip()
+    marker = "sec = "
+    start = text.find(marker)
+    if start < 0:
+        return None
+    tail = text[start + len(marker) :]
+    digits = ""
+    for char in tail:
+        if char.isdigit():
+            digits += char
+        else:
+            break
+    return f"boot-{digits}" if digits else None
+
+
+def current_boot_id() -> str | None:
+    """Identifier for the current boot, or ``None`` when it cannot be proven."""
+    system = ""
+    try:
+        system = (platform.system() or "").strip().lower()
+    except Exception:
+        system = ""
+    if system == "linux":
+        return _linux_boot_id()
+    if system == "darwin":
+        return _darwin_boot_id()
+    # An unrecognised platform yields no boot evidence rather than a guess.
+    return None
+
+
+# --- Process start time ---------------------------------------------------
+
+
+def _linux_process_start_time(pid: int) -> str | None:
+    """Field 22 of ``/proc/<pid>/stat`` — start time in clock ticks since boot.
+
+    The executable name in field 2 is parenthesised and may itself contain
+    spaces and parentheses, so the fields are located from the *last* ``)``
+    rather than by splitting the whole line.
+    """
+    try:
+        with open(f"/proc/{pid}/stat", encoding="utf-8") as handle:
+            raw = handle.read()
+    except Exception:
+        return None
+    close = raw.rfind(")")
+    if close < 0:
+        return None
+    fields = raw[close + 1 :].split()
+    # After the ')' the next field is state (field 3), so field 22 is index 19.
+    if len(fields) < 20:
+        return None
+    value = fields[19].strip()
+    return f"ticks-{value}" if value else None
+
+
+def _darwin_process_start_time(pid: int) -> str | None:
+    """macOS process start, from ``ps -o lstart=``.
+
+    ``lstart`` is the absolute wall-clock start of that pid's current
+    incarnation. Two processes reusing one pid number report different values,
+    which is exactly the discrimination reuse detection needs.
+    """
+    try:
+        completed = subprocess.run(
+            ["/bin/ps", "-o", "lstart=", "-p", str(int(pid))],
+            capture_output=True,
+            text=True,
+            timeout=5,
+            check=False,
+        )
+    except Exception:
+        return None
+    if completed.returncode != 0:
+        return None
+    value = " ".join((completed.stdout or "").split())
+    return f"lstart-{value}" if value else None
+
+
+def process_start_time(pid: int | None) -> str | None:
+    """Start-time token for *pid*'s current incarnation, or ``None``.
+
+    ``None`` is returned both when the pid does not exist and when the platform
+    cannot answer. Callers must not read either case as evidence of death — a
+    dead pid is established by the liveness probe, and this value only ever
+    *withdraws* a retirement that pid-level evidence would otherwise allow.
+    """
+    if pid is None:
+        return None
+    try:
+        numeric = int(pid)
+    except (TypeError, ValueError):
+        return None
+    if numeric <= 0:
+        return None
+    system = ""
+    try:
+        system = (platform.system() or "").strip().lower()
+    except Exception:
+        system = ""
+    if system == "linux":
+        return _linux_process_start_time(numeric)
+    if system == "darwin":
+        return _darwin_process_start_time(numeric)
+    return None
+
+
+def current_process_fencing(pid: int | None = None) -> dict[str, str | None]:
+    """The full fencing triple for *pid* (defaults to this process)."""
+    target = os.getpid() if pid is None else pid
+    return {
+        "host_id": current_host_id(),
+        "boot_id": current_boot_id(),
+        "process_start_time": process_start_time(target),
+    }

namespace_workspace_binding.py

@@ -8,8 +8,10 @@ poison workspace purity checks in another namespace.
 from __future__ import annotations
 
 import os
+import subprocess
 
 import author_mutation_worktree as amw
+import canonical_repository_root as crr
 
 ACTIVE_WORKTREE_ENV = amw.ACTIVE_WORKTREE_ENV
 AUTHOR_WORKTREE_ENV = amw.AUTHOR_WORKTREE_ENV
@@ -152,6 +154,60 @@ def resolve_namespace_workspace(
     return os.path.realpath(process_project_root), "MCP server process root (default)"
 
 
+def verify_git_common_directory_membership(
+    workspace_path: str,
+    canonical_repo_root: str,
+) -> tuple[bool, str | None]:
+    """Verify that workspace_path belongs to canonical_repo_root via git common-dir or branches containment."""
+    ws = (workspace_path or "").strip()
+    root = (canonical_repo_root or "").strip()
+    if not ws or not root:
+        return False, "empty workspace or canonical root path"
+
+    try:
+        real_ws = os.path.realpath(os.path.abspath(ws))
+        real_root = os.path.realpath(os.path.abspath(root))
+    except Exception as exc:
+        return False, f"invalid workspace or root path: {exc}"
+
+    if real_ws == real_root:
+        return True, None
+
+    if not os.path.isdir(real_ws):
+        return False, f"workspace directory '{real_ws}' does not exist"
+
+    try:
+        res = subprocess.run(
+            ["git", "-C", real_ws, "rev-parse", "--git-common-dir"],
+            capture_output=True,
+            text=True,
+            check=False,
+        )
+        if res.returncode == 0:
+            common_raw = (res.stdout or "").strip()
+            common_dir = amw._realpath_git_common_dir(real_ws, common_raw)
+            real_common = os.path.realpath(common_dir)
+            canonical_git = os.path.realpath(os.path.join(real_root, ".git"))
+
+            if real_common in (canonical_git, real_root):
+                return True, None
+            return (
+                False,
+                f"workspace '{real_ws}' git common directory '{real_common}' does not match "
+                f"canonical repository root '{real_root}' (.git at '{canonical_git}')"
+            )
+        else:
+            return (
+                False,
+                f"workspace '{real_ws}' is not a valid git repository or git rev-parse failed"
+            )
+    except Exception as exc:
+        return (
+            False,
+            f"failed to inspect git common directory for workspace '{real_ws}': {exc}"
+        )
+
+
 def resolve_namespace_mutation_context(
     *,
     role_kind: str,
@@ -163,6 +219,8 @@ def resolve_namespace_mutation_context(
     worktree: str | None = None,
     profile_name: str | None = None,
     configured_canonical_root: str | None = None,
+    expected_slug: str | None = None,
+    remote: str | None = None,
 ) -> dict:
     """Shared workspace resolution for runtime_context and mutation guards.
 
@@ -180,11 +238,41 @@ def resolve_namespace_mutation_context(
     env_map = env if env is not None else os.environ
     process_root = os.path.realpath(process_project_root)
     role = normalize_role_kind(role_kind, profile_name=profile_name)
-    configured = (configured_canonical_root or "").strip()
-    if configured:
-        canonical_root = os.path.realpath(configured)
+
+    configured_val = (configured_canonical_root or "").strip()
+    if configured_val:
+        crr_assessment = crr.assess_canonical_repository_root(
+            configured_value=configured_val,
+            source="configured_canonical_root",
+            expected_slug=expected_slug,
+            process_project_root=process_root,
+            remote=remote,
+            require_binding=True,
+        )
+        # #973 B10: a refused repository-authority mode deliberately resolves no
+        # canonical root, so downstream guards keep evaluating the install
+        # checkout rather than an identity derived through an undefined mode.
+        # ``roots_aligned`` still follows ``proven`` and stays False, and the
+        # refusal (with its reason_code) rides along in
+        # ``canonical_root_assessment`` — this is a fail-closed fallback, never a
+        # normalisation of the mode.
+        canonical_root = crr_assessment["canonical_repo_root"] or amw.resolve_canonical_repo_root(
+            process_root, process_root
+        )
+        roots_aligned = crr_assessment["proven"]
     else:
-        canonical_root = amw.resolve_canonical_repo_root(process_root, process_root)
+        crr_assessment = {
+            "proven": True,
+            "block": False,
+            "reasons": [],
+            "configured": False,
+            "canonical_repo_root": amw.resolve_canonical_repo_root(process_root, process_root),
+            "resolved_slug": None,
+            "source": None,
+            "reason_code": None,
+        }
+        canonical_root = crr_assessment["canonical_repo_root"]
+        roots_aligned = (canonical_root == process_root)
 
     durable: dict | None = None
     if role == "author":
@@ -234,7 +322,10 @@ def resolve_namespace_mutation_context(
         "ignored_bindings": demotions + (pollution.get("ignored_bindings") or []),
         "process_project_root": process_root,
         "canonical_repo_root": canonical_root,
-        "roots_aligned": canonical_root == process_root,
+        "roots_aligned": roots_aligned,
+        "canonical_root_assessment": crr_assessment,
+        "expected_slug": expected_slug,
+        "remote": remote,
     }
     if durable is not None:
         result["author_worktree_resolution"] = durable
@@ -246,6 +337,15 @@ def resolve_namespace_mutation_context(
         result["author_worktree_reasons"] = list(durable.get("reasons") or [])
         result["author_worktree_blocker_kind"] = durable.get("blocker_kind")
         result["operator_recovery"] = durable.get("operator_recovery")
+    else:
+        path_exists = os.path.exists(workspace)
+        result["path_exists"] = path_exists
+        result["in_git_worktree_list"] = (
+            amw.path_in_git_worktree_list(workspace, canonical_root)
+            if path_exists
+            else False
+        )
+        result["bound_worktree_missing"] = not path_exists
     return result
 
 
@@ -378,8 +478,8 @@ def format_namespace_workspace_binding_error(
 def assess_namespace_mutation_workspace(
     *,
     role_kind: str,
-    worktree_path: str | None,
-    worktree: str | None,
+    worktree_path: str | None = None,
+    worktree: str | None = None,
     process_project_root: str,
     env: dict[str, str] | os._Environ | None = None,
     session_lease_worktree: str | None = None,
@@ -387,6 +487,8 @@ def assess_namespace_mutation_workspace(
     profile_name: str | None = None,
     current_branch: str | None = None,
     configured_canonical_root: str | None = None,
+    expected_slug: str | None = None,
+    remote: str | None = None,
 ) -> dict:
     """Evaluate namespace workspace binding before preflight/mutation."""
     ctx = resolve_namespace_mutation_context(
@@ -399,6 +501,8 @@ def assess_namespace_mutation_workspace(
         session_lock_worktree=session_lock_worktree,
         profile_name=profile_name,
         configured_canonical_root=configured_canonical_root,
+        expected_slug=expected_slug,
+        remote=remote,
     )
     mutation_workspace = ctx["workspace_path"]
     binding_source = ctx["workspace_binding_source"]
@@ -422,6 +526,27 @@ def assess_namespace_mutation_workspace(
 
     reasons = list(metadata.get("reasons") or [])
     operator_recovery = ctx.get("operator_recovery")
+
+    crr_reasons = list(ctx.get("canonical_root_assessment", {}).get("reasons") or [])
+    if crr_reasons:
+        reasons.extend(crr_reasons)
+
+    path_exists = ctx.get("path_exists")
+    if path_exists is None:
+        path_exists = os.path.exists(mutation_workspace)
+
+    if not path_exists:
+        if role != "author":
+            reasons.append(
+                f"{role} mutation blocked: configured workspace directory '{mutation_workspace}' does not exist (nonexistent worktree)"
+            )
+    else:
+        valid_common, common_err = verify_git_common_directory_membership(
+            mutation_workspace, ctx["canonical_repo_root"]
+        )
+        if not valid_common and common_err:
+            reasons.append(common_err)
+
     if role == "author":
         # #618 durable resolution already validated existence, membership,
         # branches/, lock ownership, and traversal safety when present.
@@ -438,26 +563,27 @@ def assess_namespace_mutation_workspace(
             )
             if branches["block"]:
                 reasons.extend(branches["reasons"])
-    elif (
-        role == "reviewer"
-        and mutation_workspace == process_root
-        and not amw.is_path_under_branches(mutation_workspace, ctx["canonical_repo_root"])
-    ):
-        reasons.append(
-            f"{role} mutation blocked: workspace is the stable control checkout; "
-            f"create or reconnect to a session-owned worktree under branches/ "
-            f"or set {ROLE_WORKTREE_ENVS.get(role, ACTIVE_WORKTREE_ENV)} / "
-            f"{ACTIVE_WORKTREE_ENV}"
-        )
-    elif (
-        role in {"reviewer", "merger"}
-        and mutation_workspace != process_root
-        and not amw.is_path_under_branches(mutation_workspace, ctx["canonical_repo_root"])
-    ):
-        reasons.append(
-            f"{role} mutation blocked: workspace '{mutation_workspace}' is not under "
-            f"'{ctx['canonical_repo_root']}/branches/'"
-        )
+    elif role in {"reviewer", "merger"}:
+        if mutation_workspace == process_root and not amw.is_path_under_branches(mutation_workspace, ctx["canonical_repo_root"]):
+            reasons.append(
+                f"{role} mutation blocked: workspace is the stable control checkout; "
+                f"create or reconnect to a session-owned worktree under branches/ "
+                f"or set {ROLE_WORKTREE_ENVS.get(role, ACTIVE_WORKTREE_ENV)} / "
+                f"{ACTIVE_WORKTREE_ENV}"
+            )
+        elif mutation_workspace != process_root and not amw.is_path_under_branches(mutation_workspace, ctx["canonical_repo_root"]):
+            reasons.append(
+                f"{role} mutation blocked: workspace '{mutation_workspace}' is not under "
+                f"'{ctx['canonical_repo_root']}/branches/'"
+            )
+
+        if path_exists and amw.is_path_under_branches(mutation_workspace, ctx["canonical_repo_root"]):
+            in_list = ctx.get("in_git_worktree_list")
+            if in_list is False:
+                reasons.append(
+                    f"{role} mutation blocked: workspace '{mutation_workspace}' is under branches/ "
+                    f"but is not registered in git worktree list for '{ctx['canonical_repo_root']}'"
+                )
 
     block = bool(reasons)
     return {

post_restart_reconcile.py

@@ -475,65 +475,24 @@ def reconcile_after_restart(
             )
         )
 
-    # --- sessions (#969: dead-owner / PID-reuse lifecycle) ---------------
-    # Prefer a precomputed fleet report from the gather/apply path when present;
-    # otherwise classify pure from inventory (injectable checkers stay default).
+    # --- sessions -------------------------------------------------------
     sessions = [s for s in (inventory.get("sessions") or []) if isinstance(s, Mapping)]
-    leases_for_sessions = [
-        L for L in (inventory.get("leases") or []) if isinstance(L, Mapping)
+    orphan_sessions = [
+        s
+        for s in sessions
+        if str(s.get("status") or "").lower() == "active"
+        and s.get("pid") is not None
+        and not lease_lifecycle.is_process_alive(s.get("pid"))
     ]
-    fleet_report = inventory.get("session_fleet")
-    if isinstance(fleet_report, Mapping) and "retireable_session_ids" in fleet_report:
-        fleet_details = dict(fleet_report)
-        retireable_ids = list(fleet_details.get("retireable_session_ids") or [])
-        resolved = bool(fleet_details.get("sessions_dimension_resolved", not retireable_ids))
-    else:
-        try:
-            import session_lifecycle as _sl
-
-            client_managed = inventory.get("client_managed_session_ids")
-            cm_set = None
-            if isinstance(client_managed, (list, tuple, set, frozenset)):
-                cm_set = {str(x) for x in client_managed}
-            fleet = _sl.classify_sessions(
-                sessions,
-                leases=leases_for_sessions,
-                now=started,
-                client_managed_sessions=cm_set,
-            )
-            fleet_details = fleet.as_dict()
-            retireable_ids = list(fleet_details.get("retireable_session_ids") or [])
-            resolved = bool(fleet_details.get("sessions_dimension_resolved"))
-        except Exception as exc:  # noqa: BLE001 — fail closed to legacy signal
-            # Legacy fallback: dead-pid active rows only (pre-#969 behaviour).
-            orphan_sessions = [
-                s
-                for s in sessions
-                if str(s.get("status") or "").lower() == "active"
-                and s.get("pid") is not None
-                and not lease_lifecycle.is_process_alive(s.get("pid"))
-            ]
-            retireable_ids = [s.get("session_id") for s in orphan_sessions]
-            resolved = not orphan_sessions
-            fleet_details = {
-                "total_sessions": len(sessions),
-                "retireable_session_ids": retireable_ids,
-                "legacy_fallback": True,
-                "fallback_error": str(exc),
-            }
-
-    if not resolved and retireable_ids:
+    if orphan_sessions:
         items.append(
             _item(
                 DIM_SESSIONS,
                 ITEM_UNRESOLVED,
-                f"{len(retireable_ids)} session row(s) with dead/reused owner "
-                f"await retirement",
+                f"{len(orphan_sessions)} active session row(s) with dead owner pid",
                 details={
-                    "orphan_session_ids": retireable_ids,
-                    "retireable_session_ids": retireable_ids,
+                    "orphan_session_ids": [s.get("session_id") for s in orphan_sessions],
                     "total_sessions": len(sessions),
-                    "fleet": fleet_details,
                 },
                 follow_up=True,
             )
@@ -543,12 +502,8 @@ def reconcile_after_restart(
             _item(
                 DIM_SESSIONS,
                 ITEM_RESOLVED,
-                f"{len(sessions)} session row(s) reconciled "
-                f"(no retireable dead/reused owners)",
-                details={
-                    "total_sessions": len(sessions),
-                    "fleet": fleet_details,
-                },
+                f"{len(sessions)} session row(s) reconciled (no dead-pid orphans)",
+                details={"total_sessions": len(sessions)},
             )
         )
 

session_lifecycle.py

@@ -1,703 +0,0 @@
-"""Safe lifecycle for workflow session rows with dead or reused owners (#969).
-
-Post-restart reconciliation previously left hundreds of ``active`` session rows
-with dead owner PIDs permanently unresolved. A PID existence check alone is not
-enough: operating systems reuse PIDs, so an unrelated live process can appear to
-own a historical session.
-
-This module is the pure classification + apply core for session retirement:
-
-* distinguish live, disconnected, stale, protected, and terminal records
-* refuse retirement when a live lease or live verified owner remains
-* protect live client-managed sessions
-* detect PID reuse via process start time vs session start / heartbeat
-* terminalize confirmed-stale rows idempotently with durable audit events
-* stay safe under concurrent reconciles (CAS on status)
-
-Design mirrors ``lease_lifecycle`` / ``post_restart_reconcile``:
-
-* Pure classification accepts injectable checkers so unit tests never touch
-  real processes.
-* Apply mutations go only through :meth:`ControlPlaneDB.retire_session`.
-* Historical rows are never deleted; status moves to a terminal value and an
-  events-row records the action + reason.
-"""
-
-from __future__ import annotations
-
-import json
-import os
-import subprocess
-from dataclasses import dataclass, field
-from datetime import datetime, timedelta, timezone
-from typing import Any, Callable, Mapping, Sequence
-
-import control_plane_db as cpd
-import lease_lifecycle
-
-# Session status vocabulary.
-SESSION_STATUS_ACTIVE = "active"
-SESSION_STATUS_RETIRED = "retired"
-SESSION_STATUS_ENDED = "ended"
-SESSION_STATUS_TERMINAL = "terminal"
-
-TERMINAL_SESSION_STATUSES = frozenset(
-    {
-        SESSION_STATUS_RETIRED,
-        SESSION_STATUS_ENDED,
-        SESSION_STATUS_TERMINAL,
-        "dead",
-        "stale",
-        "orphaned",
-    }
-)
-
-# Classification outcomes for one session row.
-CLASS_LIVE = "live"
-CLASS_DISCONNECTED = "disconnected"
-CLASS_STALE = "stale"
-CLASS_PROTECTED = "protected"
-CLASS_TERMINAL = "terminal"
-
-# Stable reason codes (audit + tests).
-REASON_ALREADY_TERMINAL = "already_terminal"
-REASON_LIVE_OWNER = "live_owner"
-REASON_LIVE_LEASE = "live_lease"
-REASON_CLIENT_MANAGED_LIVE = "client_managed_live"
-REASON_DEAD_OWNER = "dead_owner"
-REASON_PID_REUSE = "pid_reuse"
-REASON_HEARTBEAT_STALE_DEAD = "heartbeat_stale_dead_owner"
-REASON_MISSING_PID = "missing_pid_no_lease"
-
-# Event type written to control-plane events table.
-EVENT_SESSION_RETIRED = "session_retired"
-
-# Default heartbeat window before a still-alive PID is treated as disconnected
-# rather than live (does not alone authorize retirement).
-DEFAULT_HEARTBEAT_STALE_SECONDS = 900
-
-# PID reuse: process start must be strictly later than session started_at by
-# more than this skew (ps lstart is second-resolution; clocks can lag).
-PID_REUSE_SKEW = timedelta(seconds=2)
-
-# Live lease freshness values that block retirement.
-_LIVE_LEASE_FRESHNESS = frozenset({"active", "live"})
-
-
-class SessionLifecycleError(RuntimeError):
-    """Fail-closed session lifecycle policy error."""
-
-
-def _utc_now() -> datetime:
-    return datetime.now(timezone.utc)
-
-
-def _parse_ts(value: str | datetime | None) -> datetime | None:
-    if value is None:
-        return None
-    if isinstance(value, datetime):
-        if value.tzinfo is None:
-            return value.replace(tzinfo=timezone.utc)
-        return value.astimezone(timezone.utc)
-    return cpd._parse_ts(str(value))
-
-
-def _ts(dt: datetime | None = None) -> str:
-    return cpd._ts(dt)
-
-
-def process_start_time(pid: int | None) -> datetime | None:
-    """Return the OS start time of *pid*, or None when it cannot be resolved.
-
-    Uses ``ps -o lstart=`` (POSIX). Failures return None rather than inventing
-    evidence — missing start time never authorizes retirement of a live PID.
-    """
-    if pid is None:
-        return None
-    try:
-        pid_i = int(pid)
-    except (TypeError, ValueError):
-        return None
-    if pid_i <= 0:
-        return None
-    try:
-        proc = subprocess.run(
-            ["ps", "-o", "lstart=", "-p", str(pid_i)],
-            capture_output=True,
-            text=True,
-            check=False,
-            timeout=2,
-        )
-    except (OSError, subprocess.SubprocessError):
-        return None
-    if proc.returncode != 0:
-        return None
-    text = (proc.stdout or "").strip()
-    if not text:
-        return None
-    try:
-        # Example: "Wed Jul 29 09:14:36 2026"
-        naive = datetime.strptime(text, "%a %b %d %H:%M:%S %Y")
-        return naive.replace(tzinfo=timezone.utc)
-    except ValueError:
-        return None
-
-
-def _lease_freshness_label(lease: Mapping[str, Any]) -> str:
-    fr = lease.get("freshness")
-    if isinstance(fr, Mapping):
-        return str(fr.get("freshness") or "").strip().lower()
-    if fr:
-        return str(fr).strip().lower()
-    # Fall back to classifying a raw lease row.
-    try:
-        return str(
-            lease_lifecycle.classify_lease_freshness(lease).get("freshness") or ""
-        ).strip().lower()
-    except Exception:  # noqa: BLE001 — pure classifier must not raise on bad rows
-        status = str(lease.get("status") or "").strip().lower()
-        return status or "unknown"
-
-
-def live_lease_session_ids(
-    leases: Sequence[Mapping[str, Any]] | None,
-    *,
-    now: datetime | None = None,
-    pid_checker: Callable[[int | None], bool] = lease_lifecycle.is_process_alive,
-) -> set[str]:
-    """Session ids that still hold a live (or ambiguous-active) workflow lease."""
-    live: set[str] = set()
-    moment = now or _utc_now()
-    for lease in leases or ():
-        if not isinstance(lease, Mapping):
-            continue
-        status = str(lease.get("status") or "").strip().lower()
-        if status and status not in {
-            lease_lifecycle.LEASE_STATUS_ACTIVE,
-            "",
-        }:
-            # Explicit terminal lease statuses never protect a session.
-            if status in {
-                lease_lifecycle.LEASE_STATUS_RELEASED,
-                lease_lifecycle.LEASE_STATUS_EXPIRED,
-                lease_lifecycle.LEASE_STATUS_ABANDONED,
-            }:
-                continue
-        freshness = _lease_freshness_label(lease)
-        if freshness in _LIVE_LEASE_FRESHNESS or freshness in {"", "unknown"}:
-            # Ambiguous active rows: re-check with authoritative classifier.
-            try:
-                fr = lease_lifecycle.classify_lease_freshness(
-                    lease, now=moment, pid_checker=pid_checker
-                )
-                freshness = str(fr.get("freshness") or "").strip().lower()
-            except Exception:  # noqa: BLE001
-                freshness = "unknown"
-        if freshness in _LIVE_LEASE_FRESHNESS:
-            sid = str(lease.get("session_id") or "").strip()
-            if sid:
-                live.add(sid)
-        elif freshness == "unknown" and status in {
-            lease_lifecycle.LEASE_STATUS_ACTIVE,
-            "",
-        }:
-            # Fail closed: active lease with unknown freshness blocks retirement.
-            sid = str(lease.get("session_id") or "").strip()
-            if sid:
-                live.add(sid)
-    return live
-
-
-@dataclass(frozen=True)
-class SessionClassification:
-    """Classification of one workflow session row."""
-
-    session_id: str
-    classification: str
-    reason: str
-    retireable: bool
-    status: str | None
-    pid: int | None
-    pid_alive: bool | None
-    pid_reused: bool
-    heartbeat_stale: bool
-    has_live_lease: bool
-    client_managed: bool
-    details: dict[str, Any] = field(default_factory=dict)
-
-    def as_dict(self) -> dict[str, Any]:
-        return {
-            "session_id": self.session_id,
-            "classification": self.classification,
-            "reason": self.reason,
-            "retireable": self.retireable,
-            "status": self.status,
-            "pid": self.pid,
-            "pid_alive": self.pid_alive,
-            "pid_reused": self.pid_reused,
-            "heartbeat_stale": self.heartbeat_stale,
-            "has_live_lease": self.has_live_lease,
-            "client_managed": self.client_managed,
-            "details": dict(self.details),
-        }
-
-
-def classify_session(
-    row: Mapping[str, Any],
-    *,
-    now: datetime | None = None,
-    pid_checker: Callable[[int | None], bool] = lease_lifecycle.is_process_alive,
-    process_start_probe: Callable[[int | None], datetime | None] = process_start_time,
-    live_lease_sessions: set[str] | frozenset[str] | None = None,
-    client_managed_sessions: set[str] | frozenset[str] | None = None,
-    heartbeat_stale_seconds: int = DEFAULT_HEARTBEAT_STALE_SECONDS,
-) -> SessionClassification:
-    """Classify one session row for retirement decisions (#969).
-
-    Rules (first match wins where noted):
-
-    1. Non-active / already-terminal status → ``terminal`` (not retireable).
-    2. Session holds a live lease → ``protected`` (never retire).
-    3. PID missing and no live lease → ``stale`` (retireable: missing owner).
-    4. PID alive + process start after session start → ``stale`` (PID reuse).
-    5. PID alive + client-managed → ``live`` protected (never retire).
-    6. PID alive + fresh heartbeat → ``live``.
-    7. PID alive + stale heartbeat → ``disconnected`` (not retireable alone).
-    8. PID dead → ``stale`` (retireable).
-    """
-    moment = now or _utc_now()
-    session_id = str(row.get("session_id") or "").strip()
-    status = str(row.get("status") or "").strip().lower() or None
-    raw_pid = row.get("pid")
-    try:
-        pid = int(raw_pid) if raw_pid is not None else None
-    except (TypeError, ValueError):
-        pid = None
-
-    client_managed = bool(
-        row.get("client_managed")
-        or row.get("is_client_managed")
-        or (
-            client_managed_sessions is not None
-            and session_id in client_managed_sessions
-        )
-    )
-    has_live_lease = bool(
-        live_lease_sessions is not None and session_id in live_lease_sessions
-    )
-
-    hb = _parse_ts(row.get("last_heartbeat_at"))
-    heartbeat_stale = bool(
-        hb is not None
-        and (moment - hb).total_seconds() > max(0, int(heartbeat_stale_seconds))
-    )
-    if hb is None and status == SESSION_STATUS_ACTIVE:
-        # No heartbeat evidence: treat as stale for liveness bookkeeping only.
-        heartbeat_stale = True
-
-    started = _parse_ts(row.get("started_at"))
-    recorded_proc_start = _parse_ts(row.get("owner_process_started_at"))
-
-    if status in TERMINAL_SESSION_STATUSES:
-        return SessionClassification(
-            session_id=session_id,
-            classification=CLASS_TERMINAL,
-            reason=REASON_ALREADY_TERMINAL,
-            retireable=False,
-            status=status,
-            pid=pid,
-            pid_alive=None,
-            pid_reused=False,
-            heartbeat_stale=heartbeat_stale,
-            has_live_lease=has_live_lease,
-            client_managed=client_managed,
-        )
-
-    if has_live_lease:
-        return SessionClassification(
-            session_id=session_id,
-            classification=CLASS_PROTECTED,
-            reason=REASON_LIVE_LEASE,
-            retireable=False,
-            status=status,
-            pid=pid,
-            pid_alive=pid_checker(pid) if pid is not None else None,
-            pid_reused=False,
-            heartbeat_stale=heartbeat_stale,
-            has_live_lease=True,
-            client_managed=client_managed,
-            details={"blocker": "live_workflow_lease"},
-        )
-
-    if pid is None:
-        return SessionClassification(
-            session_id=session_id,
-            classification=CLASS_STALE,
-            reason=REASON_MISSING_PID,
-            retireable=True,
-            status=status,
-            pid=None,
-            pid_alive=False,
-            pid_reused=False,
-            heartbeat_stale=heartbeat_stale,
-            has_live_lease=False,
-            client_managed=client_managed,
-        )
-
-    pid_alive = bool(pid_checker(pid))
-    pid_reused = False
-    proc_start: datetime | None = None
-
-    if pid_alive:
-        proc_start = recorded_proc_start or process_start_probe(pid)
-        # PID reuse: live process started after the session row itself was
-        # created. Anchor on started_at only — last_heartbeat alone is not a
-        # safe bound (synthetic inventories and long-lived processes would
-        # false-positive against a live ``ps`` probe).
-        if proc_start is not None and started is not None:
-            if proc_start > (started + PID_REUSE_SKEW):
-                pid_reused = True
-
-    if pid_reused:
-        return SessionClassification(
-            session_id=session_id,
-            classification=CLASS_STALE,
-            reason=REASON_PID_REUSE,
-            retireable=True,
-            status=status,
-            pid=pid,
-            pid_alive=True,
-            pid_reused=True,
-            heartbeat_stale=heartbeat_stale,
-            has_live_lease=False,
-            client_managed=client_managed,
-            details={
-                "process_started_at": _ts(proc_start) if proc_start else None,
-                "session_started_at": row.get("started_at"),
-                "last_heartbeat_at": row.get("last_heartbeat_at"),
-            },
-        )
-
-    if pid_alive and client_managed:
-        return SessionClassification(
-            session_id=session_id,
-            classification=CLASS_LIVE,
-            reason=REASON_CLIENT_MANAGED_LIVE,
-            retireable=False,
-            status=status,
-            pid=pid,
-            pid_alive=True,
-            pid_reused=False,
-            heartbeat_stale=heartbeat_stale,
-            has_live_lease=False,
-            client_managed=True,
-        )
-
-    if pid_alive and not heartbeat_stale:
-        return SessionClassification(
-            session_id=session_id,
-            classification=CLASS_LIVE,
-            reason=REASON_LIVE_OWNER,
-            retireable=False,
-            status=status,
-            pid=pid,
-            pid_alive=True,
-            pid_reused=False,
-            heartbeat_stale=False,
-            has_live_lease=False,
-            client_managed=client_managed,
-        )
-
-    if pid_alive and heartbeat_stale:
-        # Process still exists but has not heartbeated — disconnected, not
-        # confirmed stale. Do not retire; operator/reconnect owns next step.
-        return SessionClassification(
-            session_id=session_id,
-            classification=CLASS_DISCONNECTED,
-            reason=REASON_LIVE_OWNER,
-            retireable=False,
-            status=status,
-            pid=pid,
-            pid_alive=True,
-            pid_reused=False,
-            heartbeat_stale=True,
-            has_live_lease=False,
-            client_managed=client_managed,
-            details={"note": "alive_pid_stale_heartbeat_not_retired"},
-        )
-
-    # PID dead (or checker said not alive).
-    reason = REASON_DEAD_OWNER
-    if heartbeat_stale:
-        reason = REASON_HEARTBEAT_STALE_DEAD
-    return SessionClassification(
-        session_id=session_id,
-        classification=CLASS_STALE,
-        reason=reason,
-        retireable=True,
-        status=status,
-        pid=pid,
-        pid_alive=False,
-        pid_reused=False,
-        heartbeat_stale=heartbeat_stale,
-        has_live_lease=False,
-        client_managed=client_managed,
-    )
-
-
-@dataclass(frozen=True)
-class SessionFleetReport:
-    """Fleet-wide classification summary for reconcile + apply."""
-
-    classifications: tuple[SessionClassification, ...]
-    live_count: int
-    disconnected_count: int
-    stale_count: int
-    protected_count: int
-    terminal_count: int
-    retireable: tuple[SessionClassification, ...]
-    counts_by_reason: dict[str, int]
-
-    def as_dict(self) -> dict[str, Any]:
-        return {
-            "live_count": self.live_count,
-            "disconnected_count": self.disconnected_count,
-            "stale_count": self.stale_count,
-            "protected_count": self.protected_count,
-            "terminal_count": self.terminal_count,
-            "retireable_count": len(self.retireable),
-            "retireable_session_ids": [c.session_id for c in self.retireable],
-            "counts_by_reason": dict(self.counts_by_reason),
-            "classifications": [c.as_dict() for c in self.classifications],
-            "sessions_dimension_resolved": len(self.retireable) == 0,
-        }
-
-
-def classify_sessions(
-    sessions: Sequence[Mapping[str, Any]] | None,
-    *,
-    leases: Sequence[Mapping[str, Any]] | None = None,
-    now: datetime | None = None,
-    pid_checker: Callable[[int | None], bool] = lease_lifecycle.is_process_alive,
-    process_start_probe: Callable[[int | None], datetime | None] = process_start_time,
-    client_managed_sessions: set[str] | frozenset[str] | None = None,
-    heartbeat_stale_seconds: int = DEFAULT_HEARTBEAT_STALE_SECONDS,
-) -> SessionFleetReport:
-    """Classify a fleet of session rows against live leases (#969)."""
-    moment = now or _utc_now()
-    live_leases = live_lease_session_ids(
-        leases, now=moment, pid_checker=pid_checker
-    )
-    results: list[SessionClassification] = []
-    for row in sessions or ():
-        if not isinstance(row, Mapping):
-            continue
-        results.append(
-            classify_session(
-                row,
-                now=moment,
-                pid_checker=pid_checker,
-                process_start_probe=process_start_probe,
-                live_lease_sessions=live_leases,
-                client_managed_sessions=client_managed_sessions,
-                heartbeat_stale_seconds=heartbeat_stale_seconds,
-            )
-        )
-
-    live_count = sum(1 for c in results if c.classification == CLASS_LIVE)
-    disconnected_count = sum(
-        1 for c in results if c.classification == CLASS_DISCONNECTED
-    )
-    stale_count = sum(1 for c in results if c.classification == CLASS_STALE)
-    protected_count = sum(1 for c in results if c.classification == CLASS_PROTECTED)
-    terminal_count = sum(1 for c in results if c.classification == CLASS_TERMINAL)
-    retireable = tuple(c for c in results if c.retireable)
-    by_reason: dict[str, int] = {}
-    for c in results:
-        by_reason[c.reason] = by_reason.get(c.reason, 0) + 1
-
-    return SessionFleetReport(
-        classifications=tuple(results),
-        live_count=live_count,
-        disconnected_count=disconnected_count,
-        stale_count=stale_count,
-        protected_count=protected_count,
-        terminal_count=terminal_count,
-        retireable=retireable,
-        counts_by_reason=by_reason,
-    )
-
-
-@dataclass(frozen=True)
-class RetirementResult:
-    """Outcome of one session retirement attempt."""
-
-    session_id: str
-    outcome: str  # retired | already_terminal | skipped | blocked | missing
-    reason: str
-    prior_status: str | None = None
-    new_status: str | None = None
-    details: dict[str, Any] = field(default_factory=dict)
-
-    def as_dict(self) -> dict[str, Any]:
-        return {
-            "session_id": self.session_id,
-            "outcome": self.outcome,
-            "reason": self.reason,
-            "prior_status": self.prior_status,
-            "new_status": self.new_status,
-            "details": dict(self.details),
-        }
-
-
-def apply_session_retirements(
-    db: cpd.ControlPlaneDB,
-    report: SessionFleetReport,
-    *,
-    dry_run: bool = False,
-    actor_session_id: str | None = None,
-    now: datetime | None = None,
-) -> dict[str, Any]:
-    """Terminalize every retireable session in *report* (idempotent).
-
-    Concurrent reconciles are safe: each retirement is a CAS on
-    ``status='active'`` (or other non-terminal). A second pass that sees the
-    same session already retired records ``already_terminal`` rather than
-    duplicating audit noise beyond a single no-op outcome.
-    """
-    moment = now or _utc_now()
-    results: list[RetirementResult] = []
-    retired = 0
-    already = 0
-    blocked = 0
-    missing = 0
-
-    for classification in report.retireable:
-        if not classification.retireable:
-            continue
-        if dry_run:
-            results.append(
-                RetirementResult(
-                    session_id=classification.session_id,
-                    outcome="skipped",
-                    reason=classification.reason,
-                    prior_status=classification.status,
-                    new_status=SESSION_STATUS_RETIRED,
-                    details={"dry_run": True, **classification.details},
-                )
-            )
-            continue
-
-        applied = db.retire_session(
-            session_id=classification.session_id,
-            reason=classification.reason,
-            actor_session_id=actor_session_id,
-            details={
-                "classification": classification.classification,
-                "pid": classification.pid,
-                "pid_alive": classification.pid_alive,
-                "pid_reused": classification.pid_reused,
-                "client_managed": classification.client_managed,
-                **classification.details,
-            },
-            now=moment,
-        )
-        outcome = str(applied.get("outcome") or "missing")
-        results.append(
-            RetirementResult(
-                session_id=classification.session_id,
-                outcome=outcome,
-                reason=str(applied.get("reason") or classification.reason),
-                prior_status=applied.get("prior_status"),
-                new_status=applied.get("new_status"),
-                details=dict(applied.get("details") or {}),
-            )
-        )
-        if outcome == "retired":
-            retired += 1
-        elif outcome == "already_terminal":
-            already += 1
-        elif outcome == "blocked":
-            blocked += 1
-        else:
-            missing += 1
-
-    # Non-retireable classifications are recorded for audit completeness when
-    # dry-run lists the fleet, but apply only mutates retireable rows.
-    return {
-        "success": True,
-        "dry_run": dry_run,
-        "retired_count": retired,
-        "already_terminal_count": already,
-        "blocked_count": blocked,
-        "missing_count": missing,
-        "planned_count": len(report.retireable),
-        "results": [r.as_dict() for r in results],
-        "audit_action": EVENT_SESSION_RETIRED,
-        "actor_session_id": actor_session_id,
-        "recorded_at": _ts(moment),
-    }
-
-
-def retire_stale_sessions(
-    db: cpd.ControlPlaneDB,
-    *,
-    sessions: Sequence[Mapping[str, Any]] | None = None,
-    leases: Sequence[Mapping[str, Any]] | None = None,
-    dry_run: bool = False,
-    actor_session_id: str | None = None,
-    now: datetime | None = None,
-    pid_checker: Callable[[int | None], bool] = lease_lifecycle.is_process_alive,
-    process_start_probe: Callable[[int | None], datetime | None] = process_start_time,
-    client_managed_sessions: set[str] | frozenset[str] | None = None,
-    heartbeat_stale_seconds: int = DEFAULT_HEARTBEAT_STALE_SECONDS,
-    session_limit: int = 500,
-) -> dict[str, Any]:
-    """End-to-end plan + apply for stale session retirement (#969).
-
-    When *sessions* is omitted the control-plane DB is inventoried (active
-    rows only). Callers that already gathered inventory should pass it.
-    """
-    moment = now or _utc_now()
-    if sessions is None:
-        sessions = db.list_sessions(
-            statuses=(SESSION_STATUS_ACTIVE,),
-            limit=max(1, int(session_limit)),
-        )
-    if leases is None:
-        try:
-            leases = db.list_leases(statuses=("active",), limit=max(1, int(session_limit)))
-        except Exception:  # noqa: BLE001
-            leases = []
-
-    report = classify_sessions(
-        sessions,
-        leases=leases,
-        now=moment,
-        pid_checker=pid_checker,
-        process_start_probe=process_start_probe,
-        client_managed_sessions=client_managed_sessions,
-        heartbeat_stale_seconds=heartbeat_stale_seconds,
-    )
-    apply_result = apply_session_retirements(
-        db,
-        report,
-        dry_run=dry_run,
-        actor_session_id=actor_session_id,
-        now=moment,
-    )
-    return {
-        "success": True,
-        "fleet": report.as_dict(),
-        "apply": apply_result,
-        "sessions_dimension_resolved": (
-            report.as_dict()["sessions_dimension_resolved"]
-            if dry_run
-            else apply_result["planned_count"]
-            == (
-                apply_result["retired_count"]
-                + apply_result["already_terminal_count"]
-            )
-            and apply_result["blocked_count"] == 0
-        ),
-    }

task_capability_map.py

@@ -153,9 +153,8 @@ TASK_CAPABILITY_MAP: dict[str, dict[str, str]] = {
         "permission": "gitea.branch.push",
         "role": "author",
     },
-    # #662: post-restart reconcile is inventory + pure classification.
-    # #969: optional apply_session_cleanup retires confirmed-stale session rows
-    # through the same tool; durable follow-up Gitea issues remain separate.
+    # #662: post-restart reconcile is read-only inventory + pure classification.
+    # Durable follow-up issue creation is a separate apply path (not this task).
     "reconcile_after_restart": {
         "permission": "gitea.read",
         "role": "author",
@@ -164,14 +163,55 @@ TASK_CAPABILITY_MAP: dict[str, dict[str, str]] = {
         "permission": "gitea.read",
         "role": "author",
     },
-    # #969: explicit plan/apply path for dead-owner session retirement.
-    "retire_stale_workflow_sessions": {
+    # #978: instance-level fleet identity/health snapshot. gitea.read is the
+    # operation gate; the tool additionally restricts role_kind to
+    # controller|reconciler so author/reviewer/merger cannot use it as a
+    # mutation surface and no unrelated write permission is introduced.
+    "snapshot_instance_fleet": {
         "permission": "gitea.read",
-        "role": "author",
+        "role": "controller",
     },
-    "gitea_retire_stale_workflow_sessions": {
+    "gitea_snapshot_instance_fleet": {
         "permission": "gitea.read",
-        "role": "author",
+        "role": "controller",
+    },
+    # #980: CAS-protected retirement of conclusively stale worker
+    # registrations.
+    #
+    # Planning is observational and stays on ``gitea.read``: it opens no
+    # transaction, writes nothing, and returns only what a fleet snapshot
+    # already exposes to the same roles.
+    #
+    # Applying is a mutation and review 657 B1 established that ``gitea.read``
+    # cannot authorize it. The mutation landing in the local control-plane
+    # registry rather than in Gitea makes it *no less* a mutation, and sharing
+    # an observational permission class with plan meant any profile that could
+    # look could also destroy. It now requires its own permission,
+    # ``gitea.worker_registry.retire``, which no profile holds by default — so
+    # author, reviewer, merger, and ordinary read-only profiles fail closed on
+    # the permission itself rather than relying on the role check alone. The
+    # role restriction (controller/reconciler), runtime parity, cohort
+    # uniqueness, and the exact registry + candidate fingerprints all remain,
+    # and are now defence in depth behind the capability rather than a
+    # substitute for it.
+    #
+    # Granting the permission is a deliberate operator act in profiles.json;
+    # removing it from a profile immediately and completely revokes apply.
+    "plan_stale_worker_retirement": {
+        "permission": "gitea.read",
+        "role": "controller",
+    },
+    "gitea_plan_stale_worker_retirement": {
+        "permission": "gitea.read",
+        "role": "controller",
+    },
+    "apply_stale_worker_retirement": {
+        "permission": "gitea.worker_registry.retire",
+        "role": "controller",
+    },
+    "gitea_apply_stale_worker_retirement": {
+        "permission": "gitea.worker_registry.retire",
+        "role": "controller",
     },
     # #644: Phase 2 Web Console recovery tasks.
     "clear_stale_binding": {
@@ -396,6 +436,25 @@ TASK_CAPABILITY_MAP: dict[str, dict[str, str]] = {
         "permission": "gitea.branch.delete",
         "role": "reconciler",
     },
+    # #970: auditing missing worktree bindings is read-only; retiring one is a
+    # control-plane cleanup mutation and carries the same reconciler-only
+    # authority as any other reconciliation cleanup (review 644 B2).
+    "audit_missing_worktree_bindings": {
+        "permission": "gitea.read",
+        "role": "reconciler",
+    },
+    "gitea_audit_missing_worktree_bindings": {
+        "permission": "gitea.read",
+        "role": "reconciler",
+    },
+    "reconcile_missing_worktree_bindings": {
+        "permission": "gitea.branch.delete",
+        "role": "reconciler",
+    },
+    "gitea_reconcile_missing_worktree_bindings": {
+        "permission": "gitea.branch.delete",
+        "role": "reconciler",
+    },
     "work_issue": {
         "permission": "gitea.pr.create",
         "role": "author",
@@ -663,6 +722,8 @@ ROLE_EXCLUSIVE_TASKS: frozenset[str] = frozenset(
         "delete_branch",
         "cleanup_merged_pr_branch",
         "reconciliation_cleanup",
+        "reconcile_missing_worktree_bindings",
+        "gitea_reconcile_missing_worktree_bindings",
         "work_issue",
         "work-issue",
     }

tests/test_config_menu.py

@@ -175,8 +175,21 @@ class TestLauncherSnippets(unittest.TestCase):
     def test_only_safe_keys_no_secrets(self):
         entry = gitea_config.launcher_entry("prgs", "/cfg/profiles.json")["gitea-tools"]
         self.assertEqual(set(entry), {"command", "args", "env"})
-        self.assertEqual(set(entry["env"]), {"GITEA_MCP_CONFIG", "GITEA_MCP_PROFILE", "GITEA_CLIENT_MANAGED"})
+        # #978 B1: production launcher also injects trusted client identity.
+        required = {
+            "GITEA_MCP_CONFIG",
+            "GITEA_MCP_PROFILE",
+            "GITEA_CLIENT_MANAGED",
+            "GITEA_MCP_CLIENT",
+            "GITEA_MCP_CLIENT_INSTANCE",
+            "GITEA_MCP_INSTANCE_PROVENANCE",
+        }
+        self.assertTrue(required.issubset(set(entry["env"])), entry["env"])
         self.assertEqual(entry["env"]["GITEA_MCP_PROFILE"], "prgs")
+        self.assertTrue(
+            entry["env"]["GITEA_MCP_CLIENT_INSTANCE"].startswith("inst-"),
+            entry["env"]["GITEA_MCP_CLIENT_INSTANCE"],
+        )
         blob = json.dumps(entry).lower()
         for word in ("token", "password", "secret"):
             self.assertNotIn(word, blob)

tests/test_control_plane_db.py

@@ -37,7 +37,7 @@ class ControlPlaneDBTest(unittest.TestCase):
             rows = dict(conn.execute("SELECT key, value FROM schema_meta").fetchall())
         finally:
             conn.close()
-        self.assertEqual(rows["schema_version"], "6")
+        self.assertEqual(rows["schema_version"], "5")
         self.assertIn("DB coordinates", rows["architecture"])
         self.assertIn("bridge", rows["architecture"].lower())
 
@@ -868,7 +868,7 @@ class SessionCheckpointTest(unittest.TestCase):
             conn.close()
         self.assertIn("session_checkpoints", names)
         record = self._write()
-        self.assertEqual(record["checkpoint_schema_version"], 6)
+        self.assertEqual(record["checkpoint_schema_version"], 5)
 
     # AC2 — checkpoints written for multi-role session fixtures.
     def test_multi_role_fixtures_each_get_a_row(self) -> None:

tests/test_cross_repo_canonical_consumption.py

@@ -313,6 +313,7 @@ class TestCanonicalRootGuardBinding(_ServerHarness):
             process_project_root=self.install_root,
             remote="prgs",
             require_binding=True,
+            mode="derivation",
         )
         self.assertFalse(got.get("block"), got.get("reasons"))
         self.assertEqual(got["resolved_slug"], TARGET_SLUG)

tests/test_issue_706_canonical_repository_root.py

@@ -241,9 +241,10 @@ class TestNamespaceContextUsesConfiguredRoot(unittest.TestCase):
             process_project_root=self.install,
             env={},
             configured_canonical_root=self.target,
+            expected_slug="Scaled-Tech-Consulting/mcp-control-plane",
         )
         self.assertEqual(ctx["canonical_repo_root"], self.target)
-        self.assertFalse(ctx["roots_aligned"])
+        self.assertTrue(ctx["roots_aligned"])
 
     def test_target_worktree_is_member_of_target_root(self):
         got = nwb.amw.assess_workspace_repo_membership(
@@ -268,6 +269,7 @@ class TestNamespaceContextUsesConfiguredRoot(unittest.TestCase):
             env={},
             current_branch="feat/issue-1",
             configured_canonical_root=self.target,
+            expected_slug="Scaled-Tech-Consulting/mcp-control-plane",
         )
         self.assertFalse(assessment["block"], assessment.get("reasons"))
         self.assertEqual(assessment["canonical_repo_root"], self.target)

tests/test_issue_969_session_lifecycle.py

@@ -1,480 +0,0 @@
-"""Tests for dead-owner / PID-reuse session retirement (#969)."""
-
-from __future__ import annotations
-
-import os
-import tempfile
-import threading
-import unittest
-from concurrent.futures import ThreadPoolExecutor, as_completed
-from datetime import datetime, timedelta, timezone
-
-import control_plane_db as cpd
-import post_restart_reconcile as prr
-import session_lifecycle as sl
-
-NOW = datetime(2026, 7, 29, 12, 0, 0, tzinfo=timezone.utc)
-EARLIER = NOW - timedelta(hours=2)
-LATER = NOW + timedelta(minutes=5)
-
-
-def _session(
-    session_id: str,
-    *,
-    pid: int | None = 4242,
-    status: str = "active",
-    started_at: datetime = EARLIER,
-    last_heartbeat_at: datetime | None = None,
-    client_managed: bool = False,
-    owner_process_started_at: datetime | None = None,
-    role: str = "author",
-) -> dict:
-    hb = last_heartbeat_at if last_heartbeat_at is not None else started_at
-    row = {
-        "session_id": session_id,
-        "role": role,
-        "profile": "prgs-author",
-        "pid": pid,
-        "status": status,
-        "started_at": cpd._ts(started_at),
-        "last_heartbeat_at": cpd._ts(hb),
-        "client_managed": client_managed,
-    }
-    if owner_process_started_at is not None:
-        row["owner_process_started_at"] = cpd._ts(owner_process_started_at)
-    return row
-
-
-def _alive(pids: set[int]):
-    def _check(pid):
-        try:
-            return int(pid) in pids
-        except (TypeError, ValueError):
-            return False
-
-    return _check
-
-
-def _starts(mapping: dict[int, datetime]):
-    def _probe(pid):
-        try:
-            return mapping.get(int(pid))
-        except (TypeError, ValueError):
-            return None
-
-    return _probe
-
-
-class ClassifyDeadOwnerTests(unittest.TestCase):
-    def test_dead_owner_is_stale_and_retireable(self) -> None:
-        c = sl.classify_session(
-            _session("ghost", pid=2_000_000_000),
-            now=NOW,
-            pid_checker=_alive(set()),
-            process_start_probe=_starts({}),
-        )
-        self.assertEqual(c.classification, sl.CLASS_STALE)
-        self.assertTrue(c.retireable)
-        self.assertIn(c.reason, {sl.REASON_DEAD_OWNER, sl.REASON_HEARTBEAT_STALE_DEAD})
-
-    def test_missing_pid_is_stale(self) -> None:
-        c = sl.classify_session(
-            _session("no-pid", pid=None),
-            now=NOW,
-            pid_checker=_alive(set()),
-        )
-        self.assertEqual(c.classification, sl.CLASS_STALE)
-        self.assertEqual(c.reason, sl.REASON_MISSING_PID)
-        self.assertTrue(c.retireable)
-
-
-class PidReuseTests(unittest.TestCase):
-    def test_pid_reuse_marks_stale_not_live(self) -> None:
-        # Process with same PID started AFTER the session was recorded.
-        c = sl.classify_session(
-            _session("reused", pid=77, started_at=EARLIER, last_heartbeat_at=EARLIER),
-            now=NOW,
-            pid_checker=_alive({77}),
-            process_start_probe=_starts({77: LATER}),
-        )
-        self.assertEqual(c.classification, sl.CLASS_STALE)
-        self.assertEqual(c.reason, sl.REASON_PID_REUSE)
-        self.assertTrue(c.pid_reused)
-        self.assertTrue(c.retireable)
-
-    def test_matching_process_start_is_live(self) -> None:
-        c = sl.classify_session(
-            _session(
-                "same-proc",
-                pid=88,
-                started_at=EARLIER,
-                last_heartbeat_at=NOW - timedelta(seconds=30),
-                owner_process_started_at=EARLIER - timedelta(seconds=5),
-            ),
-            now=NOW,
-            pid_checker=_alive({88}),
-            process_start_probe=_starts({88: EARLIER - timedelta(seconds=5)}),
-        )
-        self.assertEqual(c.classification, sl.CLASS_LIVE)
-        self.assertFalse(c.retireable)
-
-
-class LiveOwnerAndLeaseTests(unittest.TestCase):
-    def test_live_owner_not_retired(self) -> None:
-        c = sl.classify_session(
-            _session(
-                "live",
-                pid=os.getpid(),
-                last_heartbeat_at=NOW - timedelta(seconds=10),
-            ),
-            now=NOW,
-            pid_checker=_alive({os.getpid()}),
-            process_start_probe=_starts({os.getpid(): EARLIER}),
-        )
-        self.assertEqual(c.classification, sl.CLASS_LIVE)
-        self.assertFalse(c.retireable)
-
-    def test_live_lease_blocks_retirement_even_if_pid_dead(self) -> None:
-        c = sl.classify_session(
-            _session("leased", pid=99999),
-            now=NOW,
-            pid_checker=_alive(set()),
-            live_lease_sessions={"leased"},
-        )
-        self.assertEqual(c.classification, sl.CLASS_PROTECTED)
-        self.assertEqual(c.reason, sl.REASON_LIVE_LEASE)
-        self.assertFalse(c.retireable)
-
-    def test_client_managed_live_never_retired(self) -> None:
-        c = sl.classify_session(
-            _session(
-                "client",
-                pid=55,
-                client_managed=True,
-                last_heartbeat_at=NOW - timedelta(seconds=5),
-            ),
-            now=NOW,
-            pid_checker=_alive({55}),
-            process_start_probe=_starts({55: EARLIER}),
-        )
-        self.assertEqual(c.classification, sl.CLASS_LIVE)
-        self.assertEqual(c.reason, sl.REASON_CLIENT_MANAGED_LIVE)
-        self.assertFalse(c.retireable)
-
-    def test_client_managed_dead_pid_is_retireable(self) -> None:
-        # Dead client process is not a live client-managed session.
-        c = sl.classify_session(
-            _session("client-dead", pid=56, client_managed=True),
-            now=NOW,
-            pid_checker=_alive(set()),
-        )
-        self.assertEqual(c.classification, sl.CLASS_STALE)
-        self.assertTrue(c.retireable)
-
-
-class TerminalAndDisconnectedTests(unittest.TestCase):
-    def test_already_terminal_not_retireable(self) -> None:
-        c = sl.classify_session(
-            _session("done", status="retired", pid=1),
-            now=NOW,
-            pid_checker=_alive(set()),
-        )
-        self.assertEqual(c.classification, sl.CLASS_TERMINAL)
-        self.assertFalse(c.retireable)
-
-    def test_alive_stale_heartbeat_is_disconnected_not_retired(self) -> None:
-        c = sl.classify_session(
-            _session(
-                "quiet",
-                pid=66,
-                last_heartbeat_at=NOW - timedelta(hours=5),
-            ),
-            now=NOW,
-            pid_checker=_alive({66}),
-            process_start_probe=_starts({66: EARLIER}),
-        )
-        self.assertEqual(c.classification, sl.CLASS_DISCONNECTED)
-        self.assertFalse(c.retireable)
-
-
-class FleetAndApplyTests(unittest.TestCase):
-    def setUp(self) -> None:
-        self._tmp = tempfile.TemporaryDirectory()
-        self.db_path = os.path.join(self._tmp.name, "cp.sqlite3")
-        self.db = cpd.ControlPlaneDB(self.db_path)
-
-    def tearDown(self) -> None:
-        self._tmp.cleanup()
-
-    def test_mixed_fleet_and_apply_retires_only_stale(self) -> None:
-        live_pid = os.getpid()
-        # Use wall-clock "now" so upsert timestamps align with classification.
-        moment = datetime.now(timezone.utc)
-        proc_start = moment - timedelta(hours=1)
-        self.db.upsert_session(
-            session_id="s-live",
-            role="author",
-            pid=live_pid,
-            status="active",
-            owner_process_started_at=cpd._ts(proc_start),
-        )
-        self.db.upsert_session(
-            session_id="s-dead", role="reviewer", pid=2_000_000_001, status="active"
-        )
-        self.db.upsert_session(
-            session_id="s-ended", role="merger", pid=3, status="ended"
-        )
-
-        sessions = self.db.list_sessions(limit=50)
-        report = sl.classify_sessions(
-            sessions,
-            leases=[],
-            now=moment,
-            pid_checker=_alive({live_pid}),
-            process_start_probe=_starts({live_pid: proc_start}),
-        )
-        self.assertGreaterEqual(report.stale_count, 1)
-        self.assertTrue(
-            any(c.session_id == "s-dead" and c.retireable for c in report.classifications)
-        )
-        self.assertTrue(
-            any(
-                c.session_id == "s-live" and not c.retireable
-                for c in report.classifications
-            )
-        )
-
-        first = sl.apply_session_retirements(
-            self.db, report, dry_run=False, actor_session_id="actor-1", now=moment
-        )
-        self.assertGreaterEqual(first["retired_count"], 1)
-        # After retirement, active list should exclude s-dead.
-        active = {
-            s["session_id"]
-            for s in self.db.list_sessions(statuses=("active",), limit=50)
-        }
-        self.assertNotIn("s-dead", active)
-        self.assertIn("s-live", active)
-
-        # Repeated cleanup is idempotent.
-        report2 = sl.classify_sessions(
-            self.db.list_sessions(limit=50),
-            leases=[],
-            now=moment,
-            pid_checker=_alive({live_pid}),
-            process_start_probe=_starts({live_pid: proc_start}),
-        )
-        second = sl.apply_session_retirements(
-            self.db, report2, dry_run=False, actor_session_id="actor-1", now=moment
-        )
-        # No double-retirement of the same row as a new mutation.
-        self.assertEqual(second["retired_count"], 0)
-
-        # Durable audit event present.
-        import sqlite3
-
-        conn = sqlite3.connect(self.db_path)
-        try:
-            events = conn.execute(
-                "SELECT event_type, message FROM events WHERE event_type = ?",
-                ("session_retired",),
-            ).fetchall()
-        finally:
-            conn.close()
-        self.assertTrue(events)
-        self.assertTrue(any("s-dead" in (m or "") for _, m in events))
-
-    def test_live_lease_blocks_db_retirement(self) -> None:
-        self.db.upsert_session(
-            session_id="s-leased", role="author", pid=2_000_000_002, status="active"
-        )
-        self.db.upsert_work_item(
-            remote="prgs",
-            org="org",
-            repo="repo",
-            kind="issue",
-            number=969,
-        )
-        result = self.db.assign_and_lease(
-            session_id="s-leased",
-            role="author",
-            remote="prgs",
-            org="org",
-            repo="repo",
-            kind="issue",
-            number=969,
-        )
-        self.assertEqual(result.outcome, "assigned")
-
-        # Inventory-style lease with explicit live freshness (authoritative for
-        # the pure classifier). DB apply also blocks on the active lease row.
-        leases = [
-            {
-                "lease_id": result.lease_id,
-                "session_id": "s-leased",
-                "status": "active",
-                "freshness": {"freshness": "active"},
-            }
-        ]
-        report = sl.classify_sessions(
-            self.db.list_sessions(statuses=("active",), limit=20),
-            leases=leases,
-            now=NOW,
-            pid_checker=_alive(set()),
-        )
-        # Classifier protects via live lease set.
-        self.assertTrue(
-            any(
-                c.session_id == "s-leased" and c.classification == sl.CLASS_PROTECTED
-                for c in report.classifications
-            )
-        )
-        apply = sl.apply_session_retirements(
-            self.db, report, dry_run=False, actor_session_id="actor", now=NOW
-        )
-        self.assertEqual(apply["retired_count"], 0)
-        active = {
-            s["session_id"]
-            for s in self.db.list_sessions(statuses=("active",), limit=20)
-        }
-        self.assertIn("s-leased", active)
-
-        # Direct DB CAS also refuses while an active lease row remains.
-        blocked = self.db.retire_session(
-            session_id="s-leased",
-            reason=sl.REASON_DEAD_OWNER,
-            actor_session_id="actor",
-            now=NOW,
-        )
-        self.assertEqual(blocked["outcome"], "blocked")
-        self.assertEqual(blocked["reason"], "live_lease")
-
-    def test_concurrent_retirement_is_idempotent(self) -> None:
-        for i in range(20):
-            self.db.upsert_session(
-                session_id=f"ghost-{i}",
-                role="author",
-                pid=3_000_000 + i,
-                status="active",
-            )
-
-        def _worker() -> dict:
-            return sl.retire_stale_sessions(
-                self.db,
-                dry_run=False,
-                actor_session_id=f"actor-{threading.get_ident()}",
-                now=NOW,
-                pid_checker=_alive(set()),
-                process_start_probe=_starts({}),
-                session_limit=100,
-            )
-
-        outcomes = []
-        with ThreadPoolExecutor(max_workers=4) as pool:
-            futs = [pool.submit(_worker) for _ in range(4)]
-            for fut in as_completed(futs):
-                outcomes.append(fut.result())
-
-        total_retired = sum(o["apply"]["retired_count"] for o in outcomes)
-        # Exactly one successful retirement per ghost row across all workers.
-        self.assertEqual(total_retired, 20)
-        active = {
-            s["session_id"]
-            for s in self.db.list_sessions(statuses=("active",), limit=100)
-        }
-        for i in range(20):
-            self.assertNotIn(f"ghost-{i}", active)
-
-
-class ReconcileIntegrationTests(unittest.TestCase):
-    def test_unresolved_until_retired_then_resolved(self) -> None:
-        inv = {
-            "inventory_complete": True,
-            "incomplete_reasons": [],
-            "service_health": {"healthy": True},
-            "clients": [{"session_id": "c1", "connected": True}],
-            "sessions": [
-                _session("ghost", pid=2_000_000_099, last_heartbeat_at=EARLIER),
-            ],
-            "leases": [],
-            "checkpoints_available": False,
-            "worktree_bindings": [],
-            "pending_mutations": [],
-            "capabilities": {"stale": False},
-            "boot_head_sha": "a" * 40,
-            "current_head_sha": "a" * 40,
-            "queue_state": {"safe_to_resume": True},
-        }
-        proof = prr.reconcile_after_restart(inv, now=NOW, mode=prr.MODE_LOG_ONLY)
-        sess = next(i for i in proof.items if i.dimension == prr.DIM_SESSIONS)
-        self.assertEqual(sess.status, prr.ITEM_UNRESOLVED)
-        self.assertIn("ghost", sess.details.get("orphan_session_ids") or [])
-
-        # After retirement inventory (no active orphans) resolves.
-        inv2 = dict(inv)
-        inv2["sessions"] = []
-        inv2["session_fleet"] = {
-            "retireable_session_ids": [],
-            "sessions_dimension_resolved": True,
-            "live_count": 0,
-            "stale_count": 0,
-        }
-        proof2 = prr.reconcile_after_restart(inv2, now=NOW, mode=prr.MODE_LOG_ONLY)
-        sess2 = next(i for i in proof2.items if i.dimension == prr.DIM_SESSIONS)
-        self.assertEqual(sess2.status, prr.ITEM_RESOLVED)
-
-    def test_legacy_orphan_key_still_populated(self) -> None:
-        inv = {
-            "inventory_complete": True,
-            "service_health": {"healthy": True},
-            "clients": [],
-            "sessions": [_session("ghost", pid=2_000_000_100)],
-            "leases": [],
-            "checkpoints_available": False,
-            "worktree_bindings": [],
-            "pending_mutations": [],
-            "capabilities": {"stale": False},
-            "boot_head_sha": "a" * 40,
-            "current_head_sha": "a" * 40,
-            "queue_state": {"safe_to_resume": True},
-        }
-        proof = prr.reconcile_after_restart(inv, now=NOW)
-        sess = next(i for i in proof.items if i.dimension == prr.DIM_SESSIONS)
-        self.assertIn("orphan_session_ids", sess.details)
-
-
-class SchemaMigrationTests(unittest.TestCase):
-    def test_lifecycle_columns_present(self) -> None:
-        with tempfile.TemporaryDirectory() as tmp:
-            path = os.path.join(tmp, "cp.sqlite3")
-            db = cpd.ControlPlaneDB(path)
-            db.upsert_session(session_id="s1", role="author", pid=1)
-            out = db.retire_session(
-                session_id="s1",
-                reason=sl.REASON_DEAD_OWNER,
-                actor_session_id="tester",
-                now=NOW,
-            )
-            self.assertEqual(out["outcome"], "retired")
-            rows = db.list_sessions(limit=5)
-            # May not appear under active filter
-            all_rows = db.list_sessions(limit=5)
-            # Re-open raw to check columns
-            import sqlite3
-
-            conn = sqlite3.connect(path)
-            try:
-                cols = {r[1] for r in conn.execute("PRAGMA table_info(sessions)")}
-                version = conn.execute(
-                    "SELECT value FROM schema_meta WHERE key='schema_version'"
-                ).fetchone()[0]
-            finally:
-                conn.close()
-            self.assertIn("retired_at", cols)
-            self.assertIn("retire_reason", cols)
-            self.assertIn("owner_process_started_at", cols)
-            self.assertEqual(version, "6")
-
-
-if __name__ == "__main__":
-    unittest.main()

tests/test_issue_973_b10_mode_contract.py

@@ -0,0 +1,739 @@
+"""Regression tests for Issue #973 blocker B10: repository-authority mode contract.
+
+Before this repair ``assess_canonical_repository_root`` never checked its
+``mode`` argument against an allowlist. Both dispatch points were permissive:
+
+* the configured-root path tested ``mode == "derivation"`` and sent every other
+  value into a catch-all ``else``, so an unsupported mode silently received
+  *validation* semantics, and
+* the single-repository default path tested ``mode == "validation"``, so an
+  unsupported mode skipped the identity comparison entirely and was strictly
+  *weaker* than validation.
+
+The measured consequence was that ``mode="invalid_mode"`` with matching expected
+and observed identities returned ``proven: True`` / ``block: False`` with no
+reasons, and that an unsupported mode passed on the default path where
+``"validation"`` correctly blocked.
+
+These tests exercise the production module directly with real git repositories —
+no patched stand-in for the function under test — and drive the production
+enforcement and mutation-context consumers rather than only the intermediate
+assessment.
+"""
+
+from __future__ import annotations
+
+import ast
+import inspect
+import os
+import subprocess
+import tempfile
+import unittest
+from unittest.mock import patch
+
+import canonical_repository_root as crr
+import gitea_config
+import gitea_mcp_server as mcp_server
+import namespace_workspace_binding as nwb
+import stable_control_runtime
+
+
+INSTALL_SLUG = "Scaled-Tech-Consulting/Gitea-Tools"
+TARGET_SLUG = "Scaled-Tech-Consulting/mcp-control-plane"
+FOREIGN_SLUG = "Someone-Else/Evil-Repo"
+
+# Explicitly supplied values that must all be refused. Omission is *not* in this
+# list: omitting the argument keeps the documented ``"validation"`` default.
+UNSUPPORTED_STRING_MODES = (
+    "invalid_mode",
+    "",
+    "validaton",          # misspelling
+    "derivaton",          # misspelling
+    "Validation",         # case variant
+    "DERIVATION",         # case variant
+    " validation",        # leading whitespace
+    "validation ",        # trailing whitespace
+    "derivation\n",       # trailing newline
+    "validation,derivation",
+)
+
+UNSUPPORTED_NON_STRING_MODES = (
+    None,
+    0,
+    1,
+    True,
+    False,
+    3.14,
+    [],
+    ["validation"],
+    {},
+    {"mode": "validation"},
+    ("validation",),
+    object(),
+)
+
+
+def _init_repo(path: str, remote_url: str, *, user: str = "Test User") -> None:
+    os.makedirs(path, exist_ok=True)
+    subprocess.run(["git", "init", "-b", "master"], cwd=path, check=True,
+                   capture_output=True)
+    subprocess.run(["git", "config", "user.email", "[email protected]"], cwd=path,
+                   check=True, capture_output=True)
+    subprocess.run(["git", "config", "user.name", user], cwd=path, check=True,
+                   capture_output=True)
+    with open(os.path.join(path, "README.md"), "w") as handle:
+        handle.write(f"{os.path.basename(path)}\n")
+    subprocess.run(["git", "add", "README.md"], cwd=path, check=True,
+                   capture_output=True)
+    subprocess.run(["git", "commit", "-m", "initial"], cwd=path, check=True,
+                   capture_output=True)
+    subprocess.run(["git", "remote", "add", "prgs", remote_url], cwd=path,
+                   check=True, capture_output=True)
+
+
+class _CanonicalRootFixture(unittest.TestCase):
+    """Real install / target / foreign git repositories, as in the #973 suite."""
+
+    def setUp(self):
+        self._tmp = tempfile.TemporaryDirectory()
+        self.tmp_dir = os.path.realpath(self._tmp.name)
+
+        self.install_root = os.path.join(self.tmp_dir, "Gitea-Tools")
+        _init_repo(
+            self.install_root,
+            "https://gitea.prgs.cc/Scaled-Tech-Consulting/Gitea-Tools.git",
+        )
+
+        self.target_root = os.path.join(self.tmp_dir, "mcp-control-plane")
+        _init_repo(
+            self.target_root,
+            "https://gitea.prgs.cc/Scaled-Tech-Consulting/mcp-control-plane.git",
+        )
+
+        self.evil_root = os.path.join(self.tmp_dir, "Evil-Repo")
+        _init_repo(
+            self.evil_root,
+            "https://gitea.prgs.cc/Someone-Else/Evil-Repo.git",
+            user="Evil User",
+        )
+
+        self.target_branches = os.path.join(self.target_root, "branches")
+        self.target_worktree = os.path.join(self.target_branches, "rev-pr-99")
+        subprocess.run(
+            ["git", "worktree", "add", "-b", "rev-pr-99", self.target_worktree],
+            cwd=self.target_root, check=True, capture_output=True,
+        )
+
+    def tearDown(self):
+        self._tmp.cleanup()
+
+    def assertRefusedForMode(self, assessment: dict, mode) -> None:
+        """Assert a fail-closed refusal attributable to *mode* and nothing else."""
+        self.assertFalse(assessment["proven"], assessment)
+        self.assertTrue(assessment["block"], assessment)
+        self.assertEqual(assessment["reason_code"], crr.DENY_UNKNOWN_MODE, assessment)
+        self.assertEqual(len(assessment["reasons"]), 1, assessment)
+        reason = assessment["reasons"][0]
+        self.assertIn("unsupported repository-authority mode", reason)
+        self.assertIn(repr(mode), reason)
+        # No trusted repository identity may be derived through an invalid mode.
+        self.assertIsNone(assessment["resolved_slug"], assessment)
+        self.assertIsNone(assessment["canonical_repo_root"], assessment)
+
+
+class TestB10UnsupportedModeIsRejected(_CanonicalRootFixture):
+    """Direct assessment tests for invalid-mode parsing."""
+
+    def test_invalid_string_with_missing_expected_identity(self):
+        """G1: blocks for the mode, not incidentally for a missing identity."""
+        assessment = crr.assess_canonical_repository_root(
+            configured_value=self.target_root,
+            source="env",
+            expected_slug=None,
+            process_project_root=self.install_root,
+            remote="prgs",
+            require_binding=True,
+            mode="invalid_mode",
+        )
+        self.assertRefusedForMode(assessment, "invalid_mode")
+        self.assertFalse(
+            any("unprovable or missing" in r for r in assessment["reasons"]),
+            "must block because the mode is unsupported, not because the expected "
+            "identity happened to be missing",
+        )
+
+    def test_invalid_string_with_matching_identities(self):
+        """G2: the contract violation — matching identities used to return proven."""
+        assessment = crr.assess_canonical_repository_root(
+            configured_value=self.target_root,
+            source="env",
+            expected_slug=TARGET_SLUG,
+            process_project_root=self.install_root,
+            remote="prgs",
+            require_binding=True,
+            mode="invalid_mode",
+        )
+        self.assertRefusedForMode(assessment, "invalid_mode")
+
+    def test_invalid_string_with_conflicting_identities(self):
+        """G3: refused for the mode, not for the incidental identity mismatch."""
+        assessment = crr.assess_canonical_repository_root(
+            configured_value=self.evil_root,
+            source="env",
+            expected_slug=INSTALL_SLUG,
+            process_project_root=self.install_root,
+            remote="prgs",
+            require_binding=True,
+            mode="invalid_mode",
+        )
+        self.assertRefusedForMode(assessment, "invalid_mode")
+        self.assertFalse(
+            any("identity mismatch" in r for r in assessment["reasons"]),
+            "the identity comparison must not have run at all",
+        )
+
+    def test_empty_string_mode(self):
+        """G4a: an explicitly supplied empty string is an unsupported value."""
+        assessment = crr.assess_canonical_repository_root(
+            configured_value=self.target_root,
+            source="env",
+            expected_slug=TARGET_SLUG,
+            process_project_root=self.install_root,
+            remote="prgs",
+            require_binding=True,
+            mode="",
+        )
+        self.assertRefusedForMode(assessment, "")
+
+    def test_explicit_none_mode(self):
+        """G4b: explicit None is refused; it is not treated as omission."""
+        assessment = crr.assess_canonical_repository_root(
+            configured_value=self.target_root,
+            source="env",
+            expected_slug=TARGET_SLUG,
+            process_project_root=self.install_root,
+            remote="prgs",
+            require_binding=True,
+            mode=None,
+        )
+        self.assertRefusedForMode(assessment, None)
+        self.assertIn("of type NoneType", assessment["reasons"][0])
+
+    def test_representative_non_string_modes(self):
+        for mode in UNSUPPORTED_NON_STRING_MODES:
+            with self.subTest(mode=repr(mode)):
+                assessment = crr.assess_canonical_repository_root(
+                    configured_value=self.target_root,
+                    source="env",
+                    expected_slug=TARGET_SLUG,
+                    process_project_root=self.install_root,
+                    remote="prgs",
+                    require_binding=True,
+                    mode=mode,
+                )
+                self.assertRefusedForMode(assessment, mode)
+                self.assertIn(
+                    f"of type {type(mode).__name__}", assessment["reasons"][0]
+                )
+
+    def test_unknown_strings_misspellings_and_whitespace_variants(self):
+        for mode in UNSUPPORTED_STRING_MODES:
+            with self.subTest(mode=repr(mode)):
+                assessment = crr.assess_canonical_repository_root(
+                    configured_value=self.target_root,
+                    source="env",
+                    expected_slug=TARGET_SLUG,
+                    process_project_root=self.install_root,
+                    remote="prgs",
+                    require_binding=True,
+                    mode=mode,
+                )
+                self.assertRefusedForMode(assessment, mode)
+
+    def test_supported_modes_are_exactly_two(self):
+        self.assertEqual(
+            crr.SUPPORTED_MODES, ("validation", "derivation")
+        )
+        self.assertIsNone(crr.unsupported_mode_reason("validation"))
+        self.assertIsNone(crr.unsupported_mode_reason("derivation"))
+        self.assertIsNotNone(crr.unsupported_mode_reason("invalid_mode"))
+
+
+class TestB10SupportedModesUnchanged(_CanonicalRootFixture):
+    """The repair must not disturb the two documented modes."""
+
+    def test_omitted_mode_defaults_to_validation(self):
+        """G4c/G4d: omission still selects validation, proven by both outcomes."""
+        matching = crr.assess_canonical_repository_root(
+            configured_value=self.target_root,
+            source="env",
+            expected_slug=TARGET_SLUG,
+            process_project_root=self.install_root,
+            remote="prgs",
+            require_binding=True,
+        )
+        self.assertTrue(matching["proven"], matching)
+        self.assertFalse(matching["block"], matching)
+        self.assertIsNone(matching["reason_code"], matching)
+        self.assertEqual(matching["resolved_slug"], TARGET_SLUG)
+
+        conflicting = crr.assess_canonical_repository_root(
+            configured_value=self.evil_root,
+            source="env",
+            expected_slug=INSTALL_SLUG,
+            process_project_root=self.install_root,
+            remote="prgs",
+            require_binding=True,
+        )
+        self.assertFalse(conflicting["proven"], conflicting)
+        self.assertTrue(conflicting["block"], conflicting)
+        self.assertIsNone(conflicting["reason_code"], conflicting)
+        self.assertTrue(
+            any("identity mismatch" in r for r in conflicting["reasons"]),
+            "omission must behave exactly like explicit validation",
+        )
+
+    def test_explicit_validation_retains_strict_behavior(self):
+        """G5a/G5b/G5c."""
+        ok = crr.assess_canonical_repository_root(
+            configured_value=self.target_root, source="env",
+            expected_slug=TARGET_SLUG, process_project_root=self.install_root,
+            remote="prgs", require_binding=True, mode="validation",
+        )
+        self.assertTrue(ok["proven"], ok)
+
+        mismatch = crr.assess_canonical_repository_root(
+            configured_value=self.evil_root, source="env",
+            expected_slug=INSTALL_SLUG, process_project_root=self.install_root,
+            remote="prgs", require_binding=True, mode="validation",
+        )
+        self.assertTrue(mismatch["block"], mismatch)
+        self.assertTrue(any("identity mismatch" in r for r in mismatch["reasons"]))
+
+        unprovable = crr.assess_canonical_repository_root(
+            configured_value=self.target_root, source="env",
+            expected_slug=None, process_project_root=self.install_root,
+            remote="prgs", require_binding=True, mode="validation",
+        )
+        self.assertTrue(unprovable["block"], unprovable)
+        self.assertTrue(
+            any("unprovable or missing" in r for r in unprovable["reasons"])
+        )
+
+    def test_explicit_derivation_retains_trusted_derivation(self):
+        """G5d: derivation still resolves identity with no expected slug."""
+        derived = crr.assess_canonical_repository_root(
+            configured_value=self.target_root, source="env",
+            expected_slug=None, process_project_root=self.install_root,
+            remote="prgs", require_binding=True, mode="derivation",
+        )
+        self.assertTrue(derived["proven"], derived)
+        self.assertFalse(derived["block"], derived)
+        self.assertIsNone(derived["reason_code"], derived)
+        self.assertEqual(derived["resolved_slug"], TARGET_SLUG)
+
+    def test_derivation_without_resolvable_remote_still_fails_closed(self):
+        no_remote = os.path.join(self.tmp_dir, "no-remote-target")
+        os.makedirs(no_remote)
+        subprocess.run(["git", "init", "-b", "master"], cwd=no_remote, check=True,
+                       capture_output=True)
+        assessment = crr.assess_canonical_repository_root(
+            configured_value=no_remote, source="env", expected_slug=None,
+            process_project_root=self.install_root, remote="prgs",
+            require_binding=True, mode="derivation",
+        )
+        self.assertTrue(assessment["block"], assessment)
+        self.assertTrue(
+            any("no resolvable" in r for r in assessment["reasons"]), assessment
+        )
+
+
+class TestB10SingleRepositoryDefaultPath(_CanonicalRootFixture):
+    """G6: on the unconfigured path an invalid mode used to be weaker than validation."""
+
+    def _assess(self, mode_kwargs: dict) -> dict:
+        return crr.assess_canonical_repository_root(
+            configured_value=None,
+            source=None,
+            expected_slug=FOREIGN_SLUG,
+            process_project_root=self.install_root,
+            remote="prgs",
+            require_binding=False,
+            **mode_kwargs,
+        )
+
+    def test_validation_blocks_a_foreign_expected_identity(self):
+        """G6b: the reference behaviour the invalid mode must not undercut."""
+        got = self._assess({"mode": "validation"})
+        self.assertFalse(got["proven"], got)
+        self.assertTrue(got["block"], got)
+        self.assertTrue(any("identity mismatch" in r for r in got["reasons"]))
+
+    def test_invalid_mode_no_longer_passes_where_validation_blocks(self):
+        """G6a: identical inputs, only the mode differs — must not fail open."""
+        invalid = self._assess({"mode": "invalid_mode"})
+        self.assertRefusedForMode(invalid, "invalid_mode")
+
+        validation = self._assess({"mode": "validation"})
+        self.assertEqual(
+            invalid["proven"], validation["proven"],
+            "an unsupported mode must never be more permissive than validation",
+        )
+        self.assertTrue(invalid["block"] and validation["block"])
+
+    def test_empty_string_mode_on_default_path(self):
+        """G6c."""
+        self.assertRefusedForMode(self._assess({"mode": ""}), "")
+
+    def test_omitted_mode_on_default_path_still_validates(self):
+        got = self._assess({})
+        self.assertFalse(got["proven"], got)
+        self.assertTrue(any("identity mismatch" in r for r in got["reasons"]), got)
+
+
+class TestB10RejectionOrdering(_CanonicalRootFixture):
+    """Refusal must precede every form of candidate-root or Git inspection."""
+
+    def test_no_git_or_identity_discovery_runs_for_an_unsupported_mode(self):
+        with patch.object(crr, "resolve_repo_toplevel") as toplevel, \
+             patch.object(crr, "repository_identity_slug") as identity:
+            assessment = crr.assess_canonical_repository_root(
+                configured_value=self.target_root,
+                source="env",
+                expected_slug=TARGET_SLUG,
+                process_project_root=self.install_root,
+                remote="prgs",
+                require_binding=True,
+                mode="invalid_mode",
+            )
+        self.assertRefusedForMode(assessment, "invalid_mode")
+        toplevel.assert_not_called()
+        identity.assert_not_called()
+
+    def test_the_same_spies_do_fire_for_a_supported_mode(self):
+        """Control: proves the previous test's assertions are not vacuous."""
+        with patch.object(crr, "resolve_repo_toplevel",
+                          wraps=crr.resolve_repo_toplevel) as toplevel, \
+             patch.object(crr, "repository_identity_slug",
+                          wraps=crr.repository_identity_slug) as identity:
+            crr.assess_canonical_repository_root(
+                configured_value=self.target_root,
+                source="env",
+                expected_slug=TARGET_SLUG,
+                process_project_root=self.install_root,
+                remote="prgs",
+                require_binding=True,
+                mode="validation",
+            )
+        toplevel.assert_called()
+        identity.assert_called()
+
+    def test_nonexistent_candidate_root_still_reports_the_mode_refusal(self):
+        """No mocks: the existence check cannot have run before the refusal."""
+        nonexistent = os.path.join(self.tmp_dir, "no-such-repository")
+        assessment = crr.assess_canonical_repository_root(
+            configured_value=nonexistent,
+            source="env",
+            expected_slug=TARGET_SLUG,
+            process_project_root=self.install_root,
+            remote="prgs",
+            require_binding=True,
+            mode="invalid_mode",
+        )
+        self.assertRefusedForMode(assessment, "invalid_mode")
+        self.assertFalse(
+            any("does not exist" in r for r in assessment["reasons"]), assessment
+        )
+
+    def test_symlinked_candidate_root_is_not_resolved_for_an_unsupported_mode(self):
+        link = os.path.join(self.tmp_dir, "target-alias")
+        os.symlink(self.target_root, link)
+        assessment = crr.assess_canonical_repository_root(
+            configured_value=link,
+            source="env",
+            expected_slug=TARGET_SLUG,
+            process_project_root=self.install_root,
+            remote="prgs",
+            require_binding=True,
+            mode="invalid_mode",
+        )
+        self.assertRefusedForMode(assessment, "invalid_mode")
+        # The refusal payload must not leak a resolved path for the candidate.
+        self.assertIsNone(assessment["canonical_repo_root"], assessment)
+
+
+class TestB10NoModeInjectionSurface(unittest.TestCase):
+    """``mode`` must not be reachable from requests, environment, or config."""
+
+    PRODUCTION_MODULES = ("gitea_mcp_server.py", "namespace_workspace_binding.py")
+
+    def _repo_root(self) -> str:
+        return os.path.dirname(os.path.dirname(os.path.abspath(__file__)))
+
+    def test_production_call_sites_only_pass_allowlisted_literal_modes(self):
+        """Valid hardcoded modes reach the correct path; nothing else is passed."""
+        seen: list[tuple[str, str | None]] = []
+        for name in self.PRODUCTION_MODULES:
+            path = os.path.join(self._repo_root(), name)
+            with open(path) as handle:
+                tree = ast.parse(handle.read())
+            for node in ast.walk(tree):
+                if not isinstance(node, ast.Call):
+                    continue
+                func = node.func
+                called = (
+                    func.attr if isinstance(func, ast.Attribute)
+                    else getattr(func, "id", None)
+                )
+                if called != "assess_canonical_repository_root":
+                    continue
+                supplied = [k for k in node.keywords if k.arg == "mode"]
+                if not supplied:
+                    seen.append((name, None))
+                    continue
+                value = supplied[0].value
+                self.assertIsInstance(
+                    value, ast.Constant,
+                    f"{name}: mode must be a literal, never a variable or expression",
+                )
+                self.assertIn(
+                    value.value, crr.SUPPORTED_MODES,
+                    f"{name}: unsupported mode literal {value.value!r}",
+                )
+                seen.append((name, value.value))
+        self.assertTrue(seen, "expected production call sites to be found")
+        # Both documented modes are exercised by production, and omission is used.
+        self.assertIn(None, [mode for _, mode in seen])
+        self.assertIn("derivation", [mode for _, mode in seen])
+
+    def test_no_public_entry_point_exposes_a_mode_parameter(self):
+        for func in (
+            nwb.resolve_namespace_mutation_context,
+            nwb.assess_namespace_mutation_workspace,
+            mcp_server._resolve_namespace_mutation_context,
+            mcp_server._enforce_canonical_repository_root,
+            mcp_server._canonical_repository_slug,
+            mcp_server._trusted_session_repository,
+            mcp_server._resolve_expected_repository_slug,
+        ):
+            with self.subTest(func=func.__name__):
+                self.assertNotIn("mode", inspect.signature(func).parameters)
+
+    def test_no_environment_key_selects_a_repository_authority_mode(self):
+        for key in gitea_config.RECOGNIZED_GITEA_ENV_KEYS:
+            self.assertNotIn(
+                "CANONICAL_REPOSITORY_MODE", key.upper(),
+                f"{key} would expose a repository-authority mode selector",
+            )
+        # An invented mode-ish variable is simply not consumed by crr.
+        env = {
+            "GITEA_CANONICAL_REPOSITORY_ROOT": "/some/path",
+            "GITEA_CANONICAL_REPOSITORY_MODE": "invalid_mode",
+        }
+        value, source = crr.configured_canonical_root(None, env)
+        self.assertEqual(value, "/some/path")
+        self.assertNotIn("mode", (source or "").lower())
+        self.assertIn(
+            "GITEA_CANONICAL_REPOSITORY_MODE",
+            gitea_config.get_unconsumed_gitea_env_overrides(env),
+            "an unknown GITEA_* key must still be rejected as unrecognised",
+        )
+
+    def test_repository_configuration_carries_no_mode_field(self):
+        profile = {
+            "canonical_repository_root": "/some/path",
+            "mode": "invalid_mode",
+        }
+        value, source = crr.configured_canonical_root(profile, {})
+        self.assertEqual(value, "/some/path")
+        self.assertEqual(source, "profile canonical_repository_root")
+
+
+class TestB10ProductionEnforcementPaths(_CanonicalRootFixture):
+    """Production enforcement, mutation-context, reviewer and merger consumers."""
+
+    def _force_invalid_mode(self):
+        """Simulate a future call site threading an unsupported mode.
+
+        The real ``assess_canonical_repository_root`` still executes — only the
+        caller-side argument is substituted — so the refusal under test is
+        produced by production code, not by a stand-in. No caller-controlled
+        ``mode`` parameter is added to any production signature to achieve this.
+        """
+        real = crr.assess_canonical_repository_root
+
+        def _wrapper(**kwargs):
+            kwargs["mode"] = "invalid_mode"
+            return real(**kwargs)
+
+        return patch.object(crr, "assess_canonical_repository_root", _wrapper)
+
+    def test_mutation_context_fails_closed_under_a_refused_mode(self):
+        with self._force_invalid_mode():
+            ctx = nwb.resolve_namespace_mutation_context(
+                role_kind="reviewer",
+                worktree_path=self.target_worktree,
+                process_project_root=self.install_root,
+                env={},
+                configured_canonical_root=self.target_root,
+                expected_slug=TARGET_SLUG,
+                remote="prgs",
+            )
+        assessment = ctx["canonical_root_assessment"]
+        self.assertFalse(ctx["roots_aligned"], ctx)
+        self.assertTrue(assessment["block"], assessment)
+        self.assertEqual(assessment["reason_code"], crr.DENY_UNKNOWN_MODE)
+        self.assertIsNone(assessment["resolved_slug"])
+        # The refused mode must not yield the candidate root as canonical.
+        self.assertNotEqual(ctx["canonical_repo_root"], self.target_root)
+        self.assertEqual(ctx["canonical_repo_root"], self.install_root)
+
+    def test_mutation_context_unchanged_for_the_supported_default(self):
+        ctx = nwb.resolve_namespace_mutation_context(
+            role_kind="reviewer",
+            worktree_path=self.target_worktree,
+            process_project_root=self.install_root,
+            env={},
+            configured_canonical_root=self.target_root,
+            expected_slug=TARGET_SLUG,
+            remote="prgs",
+        )
+        self.assertTrue(ctx["roots_aligned"], ctx)
+        self.assertEqual(ctx["canonical_repo_root"], self.target_root)
+        self.assertIsNone(ctx["canonical_root_assessment"]["reason_code"])
+
+    def test_reviewer_and_merger_authorization_fails_closed_under_a_refused_mode(self):
+        for role in ("reviewer", "merger"):
+            with self.subTest(role=role), self._force_invalid_mode():
+                assessment = nwb.assess_namespace_mutation_workspace(
+                    role_kind=role,
+                    worktree_path=self.target_worktree,
+                    worktree=None,
+                    process_project_root=self.install_root,
+                    env={},
+                    configured_canonical_root=self.target_root,
+                    expected_slug=TARGET_SLUG,
+                    remote="prgs",
+                )
+                self.assertTrue(assessment["block"], assessment)
+                self.assertTrue(
+                    any("unsupported repository-authority mode" in r
+                        for r in assessment["reasons"]),
+                    assessment,
+                )
+
+    def test_reviewer_and_merger_authorization_unchanged_for_supported_modes(self):
+        for role in ("reviewer", "merger"):
+            with self.subTest(role=role):
+                assessment = nwb.assess_namespace_mutation_workspace(
+                    role_kind=role,
+                    worktree_path=self.target_worktree,
+                    worktree=None,
+                    process_project_root=self.install_root,
+                    env={},
+                    configured_canonical_root=self.target_root,
+                    expected_slug=TARGET_SLUG,
+                    remote="prgs",
+                )
+                self.assertFalse(assessment["block"], assessment)
+
+    def test_final_mutation_gate_blocks_when_the_mode_refusal_unaligns_roots(self):
+        with self._force_invalid_mode():
+            ctx = nwb.resolve_namespace_mutation_context(
+                role_kind="reviewer",
+                worktree_path=self.target_worktree,
+                process_project_root=self.install_root,
+                env={},
+                configured_canonical_root=self.target_root,
+                expected_slug=TARGET_SLUG,
+                remote="prgs",
+            )
+        report = stable_control_runtime.build_runtime_report(
+            process_root=self.install_root,
+            checkout_branch="master",
+            runtime_head="abcdef123456",
+            active_task_workspace=ctx["workspace_path"],
+            canonical_repository_root=ctx["canonical_repo_root"],
+            workspace_roots_aligned=ctx["roots_aligned"],
+        )
+        gate = stable_control_runtime.assess_runtime_mutation_gate(report)
+        self.assertTrue(gate["block"], gate)
+
+    def test_enforce_canonical_repository_root_uses_the_validation_default(self):
+        """B8 boundary intact: a foreign configured root raises through production."""
+        with patch.object(mcp_server, "PROJECT_ROOT", self.install_root), \
+             patch.object(mcp_server, "_configured_canonical_root",
+                          return_value=(self.evil_root, "env")), \
+             patch.object(mcp_server.session_ctx, "get_session_context",
+                          return_value=None):
+            with self.assertRaises(RuntimeError) as raised:
+                mcp_server._enforce_canonical_repository_root(remote="prgs")
+        self.assertIn("identity mismatch", str(raised.exception))
+
+    def test_enforce_canonical_repository_root_refuses_a_threaded_invalid_mode(self):
+        bound = {"org": "Scaled-Tech-Consulting",
+                 "repository": "mcp-control-plane", "remote": "prgs"}
+        with patch.object(mcp_server, "PROJECT_ROOT", self.install_root), \
+             patch.object(mcp_server, "_configured_canonical_root",
+                          return_value=(self.target_root, "env")), \
+             patch.object(mcp_server.session_ctx, "get_session_context",
+                          return_value=bound), \
+             self._force_invalid_mode():
+            with self.assertRaises(RuntimeError) as raised:
+                mcp_server._enforce_canonical_repository_root(remote="prgs")
+        self.assertIn("unsupported repository-authority mode", str(raised.exception))
+
+    def test_enforce_canonical_repository_root_passes_for_a_valid_binding(self):
+        bound = {"org": "Scaled-Tech-Consulting",
+                 "repository": "mcp-control-plane", "remote": "prgs"}
+        with patch.object(mcp_server, "PROJECT_ROOT", self.install_root), \
+             patch.object(mcp_server, "_configured_canonical_root",
+                          return_value=(self.target_root, "env")), \
+             patch.object(mcp_server.session_ctx, "get_session_context",
+                          return_value=bound), \
+             patch.object(mcp_server.session_ctx, "assess_session_context",
+                          return_value={"block": False, "reasons": []}), \
+             patch.object(mcp_server, "get_profile",
+                          return_value={"profile_name": "prgs-reviewer"}):
+            mcp_server._enforce_canonical_repository_root(remote="prgs")
+
+    def test_canonical_repository_slug_derivation_unbroken(self):
+        """B9 boundary intact: legitimate cross-repository derivation still works."""
+        profile = {
+            "profile_name": "prgs-author",
+            "allowed_repositories": [TARGET_SLUG],
+            "canonical_repository_root": self.target_root,
+        }
+        with patch.object(mcp_server, "PROJECT_ROOT", self.install_root), \
+             patch.object(mcp_server.session_ctx, "get_session_context",
+                          return_value=None):
+            slug, reasons = mcp_server._canonical_repository_slug(profile, "prgs")
+            self.assertEqual(slug, TARGET_SLUG, reasons)
+            self.assertEqual(reasons, [])
+
+    def test_canonical_repository_slug_fails_closed_under_a_refused_mode(self):
+        profile = {
+            "profile_name": "prgs-author",
+            "allowed_repositories": [TARGET_SLUG],
+            "canonical_repository_root": self.target_root,
+        }
+        with patch.object(mcp_server, "PROJECT_ROOT", self.install_root), \
+             patch.object(mcp_server.session_ctx, "get_session_context",
+                          return_value=None), \
+             self._force_invalid_mode():
+            slug, reasons = mcp_server._canonical_repository_slug(profile, "prgs")
+            self.assertIsNone(slug)
+            self.assertTrue(
+                any("unsupported repository-authority mode" in r for r in reasons),
+                reasons,
+            )
+            result = mcp_server._trusted_session_repository(
+                profile, "prgs", for_mutation=True
+            )
+            self.assertIsNone(result["org"])
+            self.assertIsNone(result["repository"])
+            self.assertTrue(result["reasons"])
+
+
+if __name__ == "__main__":
+    unittest.main()

tests/test_issue_973_cross_repo_canonical_roots.py

@@ -0,0 +1,546 @@
+"""Regression tests for Issue #973: validated cross-repository canonical roots."""
+
+from __future__ import annotations
+
+import os
+import shutil
+import tempfile
+import unittest
+from unittest.mock import patch, MagicMock
+from pathlib import Path
+import subprocess
+
+import gitea_config
+import namespace_workspace_binding as nwb
+import canonical_repository_root as crr
+import stable_control_runtime
+import gitea_mcp_server as mcp_server
+
+
+class TestIssue973RecognizedEnvKeys(unittest.TestCase):
+    """Test recognized environment variable keys under #973."""
+
+    def test_recognized_gitea_env_keys(self):
+        for key in (
+            "GITEA_CANONICAL_REPOSITORY_ROOT",
+            "GITEA_REVIEWER_WORKTREE",
+            "GITEA_MERGER_WORKTREE",
+            "GITEA_MCP_SESSION_STATE_TTL_HOURS",
+        ):
+            self.assertIn(key, gitea_config.RECOGNIZED_GITEA_ENV_KEYS)
+
+    def test_get_unconsumed_gitea_env_overrides_ignores_recognized(self):
+        env = {
+            "GITEA_CANONICAL_REPOSITORY_ROOT": "/some/path",
+            "GITEA_REVIEWER_WORKTREE": "/some/reviewer/path",
+            "GITEA_MERGER_WORKTREE": "/some/merger/path",
+            "GITEA_MCP_SESSION_STATE_TTL_HOURS": "24",
+            "GITEA_UNRECOGNIZED_FOO_VAR": "bar",
+        }
+        unconsumed = gitea_config.get_unconsumed_gitea_env_overrides(env)
+        self.assertNotIn("GITEA_CANONICAL_REPOSITORY_ROOT", unconsumed)
+        self.assertNotIn("GITEA_REVIEWER_WORKTREE", unconsumed)
+        self.assertNotIn("GITEA_MERGER_WORKTREE", unconsumed)
+        self.assertNotIn("GITEA_MCP_SESSION_STATE_TTL_HOURS", unconsumed)
+        self.assertIn("GITEA_UNRECOGNIZED_FOO_VAR", unconsumed)
+
+
+class TestIssue973CrossRepoCanonicalRoots(unittest.TestCase):
+    """Test workspace binding and canonical root validation for cross-repo namespaces."""
+
+    def setUp(self):
+        self._tmp = tempfile.TemporaryDirectory()
+        self.tmp_dir = os.path.realpath(self._tmp.name)
+
+        # Create simulated installation root
+        self.install_root = os.path.join(self.tmp_dir, "Gitea-Tools")
+        os.makedirs(self.install_root)
+        subprocess.run(["git", "init", "-b", "master"], cwd=self.install_root, check=True)
+        subprocess.run(["git", "config", "user.email", "[email protected]"], cwd=self.install_root, check=True)
+        subprocess.run(["git", "config", "user.name", "Test User"], cwd=self.install_root, check=True)
+        with open(os.path.join(self.install_root, "README.md"), "w") as f:
+            f.write("install\n")
+        subprocess.run(["git", "add", "README.md"], cwd=self.install_root, check=True)
+        subprocess.run(["git", "commit", "-m", "initial"], cwd=self.install_root, check=True)
+
+        # Create simulated target repository root
+        self.target_root = os.path.join(self.tmp_dir, "mcp-control-plane")
+        os.makedirs(self.target_root)
+        subprocess.run(["git", "init", "-b", "master"], cwd=self.target_root, check=True)
+        subprocess.run(["git", "config", "user.email", "[email protected]"], cwd=self.target_root, check=True)
+        subprocess.run(["git", "config", "user.name", "Test User"], cwd=self.target_root, check=True)
+        with open(os.path.join(self.target_root, "README.md"), "w") as f:
+            f.write("target\n")
+        subprocess.run(["git", "add", "README.md"], cwd=self.target_root, check=True)
+        subprocess.run(["git", "commit", "-m", "initial"], cwd=self.target_root, check=True)
+
+        # Add remotes to simulate real git repositories with identities
+        subprocess.run(["git", "remote", "add", "prgs", "https://gitea.prgs.cc/Scaled-Tech-Consulting/Gitea-Tools.git"], cwd=self.install_root, check=True)
+        subprocess.run(["git", "remote", "add", "prgs", "https://gitea.prgs.cc/Scaled-Tech-Consulting/mcp-control-plane.git"], cwd=self.target_root, check=True)
+
+        # Create simulated foreign repository root
+        self.evil_root = os.path.join(self.tmp_dir, "Evil-Repo")
+        os.makedirs(self.evil_root)
+        subprocess.run(["git", "init", "-b", "master"], cwd=self.evil_root, check=True)
+        subprocess.run(["git", "config", "user.email", "[email protected]"], cwd=self.evil_root, check=True)
+        subprocess.run(["git", "config", "user.name", "Evil User"], cwd=self.evil_root, check=True)
+        with open(os.path.join(self.evil_root, "README.md"), "w") as f:
+            f.write("evil\n")
+        subprocess.run(["git", "add", "README.md"], cwd=self.evil_root, check=True)
+        subprocess.run(["git", "commit", "-m", "initial"], cwd=self.evil_root, check=True)
+        subprocess.run(["git", "remote", "add", "prgs", "https://gitea.prgs.cc/Someone-Else/Evil-Repo.git"], cwd=self.evil_root, check=True)
+
+        # Create branches/ directory and a valid registered worktree in target repository
+        self.target_branches = os.path.join(self.target_root, "branches")
+        self.target_worktree = os.path.join(self.target_branches, "rev-pr-99")
+        subprocess.run(["git", "worktree", "add", "-b", "rev-pr-99", self.target_worktree], cwd=self.target_root, check=True)
+
+    def tearDown(self):
+        self._tmp.cleanup()
+
+    def test_valid_same_repository_configuration(self):
+        ctx = nwb.resolve_namespace_mutation_context(
+            role_kind="reviewer",
+            worktree_path=None,
+            process_project_root=self.install_root,
+            env={},
+            configured_canonical_root=None,
+        )
+        self.assertEqual(ctx["canonical_repo_root"], self.install_root)
+        self.assertTrue(ctx["roots_aligned"])
+        self.assertTrue(ctx["canonical_root_assessment"]["proven"])
+
+    def test_valid_cross_repo_canonical_root(self):
+        ctx = nwb.resolve_namespace_mutation_context(
+            role_kind="reviewer",
+            worktree_path=self.target_worktree,
+            process_project_root=self.install_root,
+            env={},
+            configured_canonical_root=self.target_root,
+            expected_slug="Scaled-Tech-Consulting/mcp-control-plane",
+        )
+        self.assertEqual(ctx["canonical_repo_root"], self.target_root)
+        self.assertTrue(ctx["roots_aligned"])
+        self.assertTrue(ctx["canonical_root_assessment"]["proven"])
+
+    def test_expected_repository_identity_match(self):
+        assessment = crr.assess_canonical_repository_root(
+            configured_value=self.target_root,
+            source="test",
+            expected_slug="Scaled-Tech-Consulting/mcp-control-plane",
+            process_project_root=self.install_root,
+            remote="prgs",
+        )
+        self.assertTrue(assessment["proven"])
+        self.assertFalse(assessment["block"])
+
+    def test_foreign_repository_identity_mismatch(self):
+        assessment = crr.assess_canonical_repository_root(
+            configured_value=self.evil_root,
+            source="test",
+            expected_slug="Scaled-Tech-Consulting/Gitea-Tools",
+            process_project_root=self.install_root,
+            remote="prgs",
+        )
+        self.assertFalse(assessment["proven"])
+        self.assertTrue(assessment["block"])
+        self.assertTrue(any("identity mismatch" in r for r in assessment["reasons"]))
+
+    def test_native_repository_binding_mismatch(self):
+        ctx = nwb.resolve_namespace_mutation_context(
+            role_kind="reviewer",
+            worktree_path=self.target_worktree,
+            process_project_root=self.install_root,
+            env={},
+            configured_canonical_root=self.evil_root,
+            expected_slug="Scaled-Tech-Consulting/Gitea-Tools",
+            remote="prgs",
+        )
+        self.assertFalse(ctx["roots_aligned"])
+        self.assertFalse(ctx["canonical_root_assessment"]["proven"])
+        self.assertTrue(any("identity mismatch" in r for r in ctx["canonical_root_assessment"]["reasons"]))
+
+    def test_unpatched_foreign_configured_root_derives_expected_from_process_root_and_blocks(self):
+        """B8: Production path test where foreign configured root cannot self-authorize."""
+        with patch.object(mcp_server, "PROJECT_ROOT", self.install_root), \
+             patch.object(mcp_server, "_configured_canonical_root", return_value=(self.evil_root, "env")):
+            expected_slug = mcp_server._resolve_expected_repository_slug("prgs")
+            self.assertEqual(expected_slug, "Scaled-Tech-Consulting/Gitea-Tools")
+            assessment = crr.assess_canonical_repository_root(
+                configured_value=self.evil_root,
+                source="env",
+                expected_slug=expected_slug,
+                process_project_root=self.install_root,
+                remote="prgs",
+                require_binding=True,
+            )
+            self.assertFalse(assessment["proven"])
+            self.assertTrue(assessment["block"])
+            self.assertEqual(assessment["resolved_slug"], "Someone-Else/Evil-Repo")
+            self.assertTrue(any("identity mismatch" in r for r in assessment["reasons"]))
+
+    def test_unpatched_valid_cross_repo_matching_session_context(self):
+        """B8: Valid cross-repo namespace matches when session context is bound to target repo."""
+        bound_ctx = {"org": "Scaled-Tech-Consulting", "repository": "mcp-control-plane", "remote": "prgs"}
+        with patch.object(mcp_server, "PROJECT_ROOT", self.install_root), \
+             patch.object(mcp_server.session_ctx, "get_session_context", return_value=bound_ctx):
+            expected_slug = mcp_server._resolve_expected_repository_slug("prgs")
+            self.assertEqual(expected_slug, "Scaled-Tech-Consulting/mcp-control-plane")
+            assessment = crr.assess_canonical_repository_root(
+                configured_value=self.target_root,
+                source="env",
+                expected_slug=expected_slug,
+                process_project_root=self.install_root,
+                remote="prgs",
+                require_binding=True,
+            )
+            self.assertTrue(assessment["proven"])
+            self.assertFalse(assessment["block"])
+            self.assertEqual(assessment["resolved_slug"], "Scaled-Tech-Consulting/mcp-control-plane")
+
+    def test_unprovable_expected_identity_fails_closed(self):
+        """B8: If expected repository identity is unprovable for a configured root, fail closed."""
+        no_remote_root = os.path.join(self.tmp_dir, "no-remote-process-root")
+        os.makedirs(no_remote_root)
+        subprocess.run(["git", "init", "-b", "master"], cwd=no_remote_root, check=True)
+        with patch.object(mcp_server, "PROJECT_ROOT", no_remote_root), \
+             patch.object(mcp_server.session_ctx, "get_session_context", return_value=None):
+            expected_slug = mcp_server._resolve_expected_repository_slug("prgs")
+            self.assertIsNone(expected_slug)
+            assessment = crr.assess_canonical_repository_root(
+                configured_value=self.target_root,
+                source="env",
+                expected_slug=expected_slug,
+                process_project_root=no_remote_root,
+                remote="prgs",
+                require_binding=True,
+            )
+            self.assertFalse(assessment["proven"])
+            self.assertTrue(assessment["block"])
+            self.assertTrue(any("unprovable or missing" in r for r in assessment["reasons"]))
+
+    def test_missing_canonical_root(self):
+        ctx = nwb.resolve_namespace_mutation_context(
+            role_kind="author",
+            worktree_path=None,
+            process_project_root=self.install_root,
+            env={},
+            configured_canonical_root="",
+        )
+        self.assertEqual(ctx["canonical_repo_root"], self.install_root)
+        self.assertTrue(ctx["roots_aligned"])
+
+    def test_nonexistent_configured_canonical_root(self):
+        nonexistent = os.path.join(self.tmp_dir, "nonexistent-repo")
+        ctx = nwb.resolve_namespace_mutation_context(
+            role_kind="reviewer",
+            worktree_path=self.target_worktree,
+            process_project_root=self.install_root,
+            env={},
+            configured_canonical_root=nonexistent,
+        )
+        self.assertFalse(ctx["roots_aligned"])
+        self.assertFalse(ctx["canonical_root_assessment"]["proven"])
+        self.assertTrue(any("does not exist" in r for r in ctx["canonical_root_assessment"]["reasons"]))
+
+    def test_non_git_configured_canonical_root(self):
+        non_git = os.path.join(self.tmp_dir, "non-git-dir")
+        os.makedirs(non_git)
+        ctx = nwb.resolve_namespace_mutation_context(
+            role_kind="reviewer",
+            worktree_path=self.target_worktree,
+            process_project_root=self.install_root,
+            env={},
+            configured_canonical_root=non_git,
+        )
+        self.assertFalse(ctx["roots_aligned"])
+        self.assertFalse(ctx["canonical_root_assessment"]["proven"])
+        self.assertTrue(any("not a git repository" in r for r in ctx["canonical_root_assessment"]["reasons"]))
+
+    def test_git_common_directory_membership_matching(self):
+        valid, err = nwb.verify_git_common_directory_membership(
+            self.target_worktree, self.target_root
+        )
+        self.assertTrue(valid, err)
+        self.assertIsNone(err)
+
+    def test_foreign_git_common_directory(self):
+        # Foreign worktree created under install_root
+        install_branches = os.path.join(self.install_root, "branches")
+        foreign_wt = os.path.join(install_branches, "foreign-wt")
+        subprocess.run(["git", "worktree", "add", "-b", "foreign-wt", foreign_wt], cwd=self.install_root, check=True)
+
+        valid, err = nwb.verify_git_common_directory_membership(
+            foreign_wt, self.target_root
+        )
+        self.assertFalse(valid)
+        self.assertIn("does not match", err)
+
+    def test_normalized_path_aliases(self):
+        alias_path = self.target_worktree + "/../rev-pr-99/./"
+        valid, err = nwb.verify_git_common_directory_membership(
+            alias_path, self.target_root
+        )
+        self.assertTrue(valid, err)
+
+    def test_safe_symlink_identity(self):
+        link_path = os.path.join(self.target_branches, "symlink-rev-99")
+        try:
+            os.symlink(self.target_worktree, link_path)
+            valid, err = nwb.verify_git_common_directory_membership(
+                link_path, self.target_root
+            )
+            self.assertTrue(valid, err)
+        finally:
+            if os.path.exists(link_path):
+                os.unlink(link_path)
+
+    def test_symlink_escape_or_foreign_alias(self):
+        outside_dir = os.path.join(self.tmp_dir, "outside-target")
+        os.makedirs(outside_dir)
+        link_escape = os.path.join(self.target_branches, "escape-link")
+        try:
+            os.symlink(outside_dir, link_escape)
+            assessment = nwb.assess_namespace_mutation_workspace(
+                role_kind="reviewer",
+                worktree_path=link_escape,
+                worktree=None,
+                process_project_root=self.install_root,
+                env={},
+                configured_canonical_root=self.target_root,
+                expected_slug="Scaled-Tech-Consulting/mcp-control-plane",
+            )
+            self.assertTrue(assessment["block"])
+        finally:
+            if os.path.exists(link_escape):
+                os.unlink(link_escape)
+
+    def test_reviewer_worktree_registered_and_valid(self):
+        assessment = nwb.assess_namespace_mutation_workspace(
+            role_kind="reviewer",
+            worktree_path=self.target_worktree,
+            worktree=None,
+            process_project_root=self.install_root,
+            env={},
+            configured_canonical_root=self.target_root,
+            expected_slug="Scaled-Tech-Consulting/mcp-control-plane",
+        )
+        self.assertFalse(assessment["block"])
+
+    def test_reviewer_worktree_unregistered_blocks(self):
+        unreg_wt = os.path.join(self.target_branches, "unregistered-reviewer")
+        os.makedirs(unreg_wt)
+        assessment = nwb.assess_namespace_mutation_workspace(
+            role_kind="reviewer",
+            worktree_path=unreg_wt,
+            worktree=None,
+            process_project_root=self.install_root,
+            env={},
+            configured_canonical_root=self.target_root,
+            expected_slug="Scaled-Tech-Consulting/mcp-control-plane",
+        )
+        self.assertTrue(assessment["block"])
+        self.assertTrue(any("is not registered in git worktree list" in r for r in assessment["reasons"]))
+
+    def test_merger_worktree_registered_and_valid(self):
+        assessment = nwb.assess_namespace_mutation_workspace(
+            role_kind="merger",
+            worktree_path=self.target_worktree,
+            worktree=None,
+            process_project_root=self.install_root,
+            env={},
+            configured_canonical_root=self.target_root,
+            expected_slug="Scaled-Tech-Consulting/mcp-control-plane",
+        )
+        self.assertFalse(assessment["block"])
+
+    def test_merger_worktree_unregistered_blocks(self):
+        unreg_wt = os.path.join(self.target_branches, "unregistered-merger")
+        os.makedirs(unreg_wt)
+        assessment = nwb.assess_namespace_mutation_workspace(
+            role_kind="merger",
+            worktree_path=unreg_wt,
+            worktree=None,
+            process_project_root=self.install_root,
+            env={},
+            configured_canonical_root=self.target_root,
+            expected_slug="Scaled-Tech-Consulting/mcp-control-plane",
+        )
+        self.assertTrue(assessment["block"])
+        self.assertTrue(any("is not registered in git worktree list" in r for r in assessment["reasons"]))
+
+    def test_reviewer_or_merger_worktree_outside_branches_blocks(self):
+        outside_wt = os.path.join(self.target_root, "outside_branches_wt")
+        os.makedirs(outside_wt)
+        for r_kind in ("reviewer", "merger"):
+            assessment = nwb.assess_namespace_mutation_workspace(
+                role_kind=r_kind,
+                worktree_path=outside_wt,
+                worktree=None,
+                process_project_root=self.install_root,
+                env={},
+                configured_canonical_root=self.target_root,
+                expected_slug="Scaled-Tech-Consulting/mcp-control-plane",
+            )
+            self.assertTrue(assessment["block"])
+            self.assertTrue(any("is not under" in r for r in assessment["reasons"]))
+
+    def test_nonexistent_reviewer_or_merger_worktree_blocks(self):
+        nonexistent_wt = os.path.join(self.target_branches, "nonexistent-wt")
+        for r_kind in ("reviewer", "merger"):
+            assessment = nwb.assess_namespace_mutation_workspace(
+                role_kind=r_kind,
+                worktree_path=nonexistent_wt,
+                worktree=None,
+                process_project_root=self.install_root,
+                env={},
+                configured_canonical_root=self.target_root,
+                expected_slug="Scaled-Tech-Consulting/mcp-control-plane",
+            )
+            self.assertTrue(assessment["block"])
+            self.assertTrue(any("does not exist" in r for r in assessment["reasons"]))
+
+    def test_safe_and_unsafe_mutation_alignment_outcomes(self):
+        # Safe alignment (same repo)
+        report_safe = stable_control_runtime.build_runtime_report(
+            process_root=self.install_root,
+            checkout_branch="master",
+            runtime_head="abcdef123456",
+            active_task_workspace=self.install_root,
+            canonical_repository_root=self.install_root,
+            workspace_roots_aligned=True,
+        )
+        gate_safe = stable_control_runtime.assess_runtime_mutation_gate(report_safe)
+        self.assertFalse(gate_safe["block"])
+
+        # Unsafe alignment
+        report_unsafe = stable_control_runtime.build_runtime_report(
+            process_root=self.install_root,
+            checkout_branch="master",
+            runtime_head="abcdef123456",
+            active_task_workspace=self.target_worktree,
+            canonical_repository_root=self.target_root,
+            workspace_roots_aligned=False,
+        )
+        gate_unsafe = stable_control_runtime.assess_runtime_mutation_gate(report_unsafe)
+        self.assertTrue(gate_unsafe["block"])
+
+    def test_reviewer_lease_lifecycle_production_path(self):
+        """B5: Automated regression for reviewer lease acquire and release through production path."""
+        mock_whoami = {
+            "authenticated": True,
+            "username": "sysadmin",
+            "remote": "prgs",
+            "profile": {
+                "profile_name": "prgs-reviewer",
+                "role": "reviewer",
+                "role_kind": "reviewer",
+                "allowed_operations": ["gitea.read", "gitea.pr.comment", "gitea.pr.approve", "gitea.pr.request_changes"],
+                "forbidden_operations": [],
+            },
+        }
+
+        def mock_api_request(method, url, auth=None, json_data=None):
+            if method == "GET":
+                return {"number": 99, "head": {"sha": "abc1234"}, "state": "open", "merged": False, "merged_at": None}
+            elif method == "POST":
+                return {"id": 9999, "body": (json_data or {}).get("body", "")}
+            return {}
+
+        with patch.object(mcp_server, "gitea_whoami", return_value=mock_whoami), \
+             patch.object(mcp_server, "get_profile", return_value=mock_whoami["profile"]), \
+             patch.object(mcp_server, "_effective_workspace_role", return_value="reviewer"), \
+             patch.object(mcp_server, "_configured_canonical_root", return_value=(self.target_root, "env")), \
+             patch.object(mcp_server, "_reviewer_session_worktree", return_value=self.target_worktree), \
+             patch.object(mcp_server, "_auth", return_value="token mock-token"), \
+             patch.object(mcp_server, "_fetch_pr_comments", return_value=[]), \
+             patch.object(mcp_server, "api_request", side_effect=mock_api_request), \
+             patch("reviewer_pr_lease.assess_acquire_lease", return_value={"acquire_allowed": True, "reasons": [], "lease_body": "<!-- LEASE -->"}), \
+             patch("reviewer_pr_lease.find_active_reviewer_lease", return_value={"session_id": "sid-123", "reviewer": "sysadmin"}), \
+             patch("reviewer_pr_lease.get_session_lease", return_value={"session_id": "sid-123", "reviewer": "sysadmin"}), \
+             patch("reviewer_pr_lease.clear_session_lease") as mock_clear:
+
+            acq_res = mcp_server.gitea_acquire_reviewer_pr_lease(
+                pr_number=99,
+                remote="prgs",
+                worktree=self.target_worktree,
+                org="Scaled-Tech-Consulting",
+                repo="mcp-control-plane",
+            )
+            self.assertTrue(acq_res.get("success"), acq_res)
+
+            rel_res = mcp_server.gitea_release_reviewer_pr_lease(
+                pr_number=99,
+                worktree=self.target_worktree,
+                remote="prgs",
+                org="Scaled-Tech-Consulting",
+                repo="mcp-control-plane",
+            )
+            self.assertTrue(rel_res.get("success"), rel_res)
+            mock_clear.assert_called_once()
+
+    def test_unbound_session_derives_and_retains_configured_target_repository(self):
+        """B9: Unbound session with a legitimate configured target root derives and retains target repository."""
+        prof = {
+            "profile_name": "prgs-author",
+            "allowed_repositories": ["Scaled-Tech-Consulting/mcp-control-plane"],
+            "canonical_repository_root": self.target_root,
+        }
+        with patch.object(mcp_server, "PROJECT_ROOT", self.install_root), \
+             patch.object(mcp_server.session_ctx, "get_session_context", return_value=None):
+            slug, reasons = mcp_server._canonical_repository_slug(prof, "prgs")
+            self.assertEqual(slug, "Scaled-Tech-Consulting/mcp-control-plane", f"reasons: {reasons}")
+            self.assertEqual(reasons, [])
+            res = mcp_server._trusted_session_repository(prof, "prgs")
+            self.assertEqual(res["org"], "Scaled-Tech-Consulting")
+            self.assertEqual(res["repository"], "mcp-control-plane")
+
+    def test_process_root_a_configured_target_b_retains_b(self):
+        """B9: Process root A (Gitea-Tools) and configured target B (mcp-control-plane) keep B as expected identity when bound."""
+        bound_ctx = {"org": "Scaled-Tech-Consulting", "repository": "mcp-control-plane", "remote": "prgs"}
+        with patch.object(mcp_server, "PROJECT_ROOT", self.install_root), \
+             patch.object(mcp_server.session_ctx, "get_session_context", return_value=bound_ctx):
+            expected = mcp_server._resolve_expected_repository_slug("prgs")
+            self.assertEqual(expected, "Scaled-Tech-Consulting/mcp-control-plane")
+
+    def test_request_supplied_repository_c_cannot_replace_derived_or_bound_repository_b(self):
+        """B9: Request-supplied repository C (Timesheet) cannot replace bound repository B (mcp-control-plane)."""
+        bound_ctx = {"org": "Scaled-Tech-Consulting", "repository": "mcp-control-plane", "remote": "prgs"}
+        with patch.object(mcp_server, "PROJECT_ROOT", self.install_root), \
+             patch.object(mcp_server.session_ctx, "get_session_context", return_value=bound_ctx):
+            expected = mcp_server._resolve_expected_repository_slug("prgs", org="Scaled-Tech-Consulting", repo="Timesheet")
+            self.assertEqual(expected, "Scaled-Tech-Consulting/mcp-control-plane")
+
+    def test_known_expected_repo_a_plus_candidate_root_b_fails_closed(self):
+        """B9: Known expected repository A plus candidate root B fails closed in validation mode."""
+        assessment = crr.assess_canonical_repository_root(
+            configured_value=self.target_root,
+            source="env",
+            expected_slug="Scaled-Tech-Consulting/Gitea-Tools",
+            process_project_root=self.install_root,
+            remote="prgs",
+            require_binding=True,
+            mode="validation",
+        )
+        self.assertFalse(assessment["proven"])
+        self.assertTrue(assessment["block"])
+        self.assertTrue(any("identity mismatch" in r for r in assessment["reasons"]))
+
+    def test_validation_mode_with_unprovable_expected_identity_fails_closed(self):
+        """B9: Validation mode with expected_slug=None and require_binding=True fails closed."""
+        assessment = crr.assess_canonical_repository_root(
+            configured_value=self.target_root,
+            source="env",
+            expected_slug=None,
+            process_project_root=self.install_root,
+            remote="prgs",
+            require_binding=True,
+            mode="validation",
+        )
+        self.assertFalse(assessment["proven"])
+        self.assertTrue(assessment["block"])
+        self.assertTrue(any("unprovable or missing" in r for r in assessment["reasons"]))
+
+
+if __name__ == "__main__":
+    unittest.main()

tests/test_namespace_workspace_binding.py

@@ -245,10 +245,15 @@ class TestNamespaceWorkspaceIntegration(unittest.TestCase):
     def test_pr487_style_merge_binds_clean_merger_workspace(
         self, _exists, _isdir, mock_run
     ):
-        mock_run.return_value = MagicMock(returncode=0, stdout=f"{CONTROL_ROOT}/.git\n")
+        def mock_git(cmd, *args, **kwargs):
+            if "rev-parse" in cmd:
+                return MagicMock(returncode=0, stdout=f"{CONTROL_ROOT}/.git\n")
+            return MagicMock(returncode=0, stdout=f"worktree {CONTROL_ROOT}\nworktree {MERGER_CLEAN}\n")
+        mock_run.side_effect = mock_git
         os.environ[nwb.AUTHOR_WORKTREE_ENV] = AUTHOR_DIRTY
+        os.environ[nwb.MERGER_WORKTREE_ENV] = MERGER_CLEAN
         srv._preflight_resolved_role = "reviewer"
         with mock.patch.object(srv, "PROJECT_ROOT", MCP_PROCESS_ROOT):
             with mock.patch("gitea_mcp_server.get_profile", return_value=self._merger_profile()):
                 resolved = srv._verify_role_mutation_workspace("prgs")
-        self.assertEqual(resolved, os.path.realpath(MCP_PROCESS_ROOT))
+        self.assertEqual(resolved, os.path.realpath(MERGER_CLEAN))

tests/test_task_capability_role_invariants.py

@@ -153,6 +153,12 @@ EXPECTED_ROLE_EXCLUSIVE_TASKS = frozenset(
         "delete_branch",
         "cleanup_merged_pr_branch",
         "reconciliation_cleanup",
+        # #970 review 644 B2: retiring a missing worktree binding is a
+        # control-plane cleanup mutation, so it carries the same reconciler-only
+        # authority as every other reconciliation cleanup. Permission alone must
+        # not authorize it.
+        "reconcile_missing_worktree_bindings",
+        "gitea_reconcile_missing_worktree_bindings",
         "work_issue",
         "work-issue",
     }