Scaled-Tech-Consulting/Gitea-Tools
2
0
Fork 0
Code Issues 73 Pull Requests 9 Actions Packages Projects Releases Wiki Activity

Reviewer mutation tools must use the active branches worktree, not stale MCP server process root #503

New Issue
Closed
opened 2026-07-08 02:25:58 -05:00 by jcwalker3 · 2 comments
jcwalker3 commented 2026-07-08 02:25:58 -05:00
Owner
Copy Link
Copy Source

Problem

Reviewer mutation tools are failing closed under #274/#475 because they resolve their mutation workspace from the MCP server process root instead of the active review worktree.

In the observed PR #487 session, get_runtime_context reported:

  • MCP server process root: /Users/jasonwalker/Development/Gitea-Tools
  • Prepared clean review worktree: /Users/jasonwalker/Development/Gitea-Tools/branches/review-pr487-submit
  • Preflight with worktree path: passed
  • Mutation tools: still failed because they resolved workspace to the root checkout

The worktree_path argument appears to affect read-only/preflight reporting and lease metadata, but does not redirect mutation execution. Tools including lease acquisition, dry-run review, final decision marking, and review submission either ignore the worktree argument for execution or do not accept one at all.

Impact

  • Valid clean branches/ worktrees cannot submit formal reviews.
  • Reviewers can validate PRs but cannot post approvals/comments.
  • Sessions repeatedly fail despite correct user behavior.
  • The root checkout remains a hard blocker whenever the MCP server was launched from the control checkout.
  • This creates false PR blockers and noisy handoffs.

Required behavior

  1. Mutation tools must resolve workspace from the active task worktree, not only from MCP server process root.
  2. If mutation tools intentionally require server launch from the worktree, get_runtime_context and preflight must fail early with a clear message before any review attempt.
  3. Passing a worktree path must either:
    • fully redirect mutation workspace validation, or
    • be rejected as metadata-only with a clear error.
  4. dry_run_pr_review, mark_final_review_decision, and submit_pr_review need an explicit workspace/worktree binding check or must inherit the validated active task workspace.
  5. The error message should say exactly how to fix it:
    • relaunch/reconnect MCP server from the branches worktree, or
    • set the required environment variable, or
    • use the supported workspace-scoping mechanism.

Acceptance criteria

  1. Add tests proving reviewer mutations succeed when the active workspace is a clean branches/ worktree.
  2. Add tests proving mutations fail closed when the MCP server is rooted at the control checkout.
  3. Add tests proving preflight and mutation tools agree on the same workspace root.
  4. Add tests proving a metadata-only worktree_path cannot produce a misleading preflight_ready: true state when mutation tools will still fail.
  5. Add a canonical handoff/error message for stale MCP server workspace binding.
  6. Confirm root checkout remains protected and cannot be used for author/reviewer mutations.

Non-goals

  • Do not weaken #274/#475 branches-only mutation guards.
  • Do not allow reviewer mutations from the stable control checkout.
  • Do not bypass existing capability or lease gates.

Related

  • #274 — branches-only author mutation guard
  • #475 — hard guard preventing root checkout drift
  • #460 — branches-only guard vs runtime_context worktree resolution
  • #487 — observed failure session (if applicable)
## Problem Reviewer mutation tools are failing closed under #274/#475 because they resolve their mutation workspace from the MCP server process root instead of the active review worktree. In the observed PR #487 session, `get_runtime_context` reported: - MCP server process root: `/Users/jasonwalker/Development/Gitea-Tools` - Prepared clean review worktree: `/Users/jasonwalker/Development/Gitea-Tools/branches/review-pr487-submit` - Preflight with worktree path: passed - Mutation tools: still failed because they resolved workspace to the root checkout The `worktree_path` argument appears to affect read-only/preflight reporting and lease metadata, but does not redirect mutation execution. Tools including lease acquisition, dry-run review, final decision marking, and review submission either ignore the worktree argument for execution or do not accept one at all. ## Impact - Valid clean `branches/` worktrees cannot submit formal reviews. - Reviewers can validate PRs but cannot post approvals/comments. - Sessions repeatedly fail despite correct user behavior. - The root checkout remains a hard blocker whenever the MCP server was launched from the control checkout. - This creates false PR blockers and noisy handoffs. ## Required behavior 1. Mutation tools must resolve workspace from the active task worktree, not only from MCP server process root. 2. If mutation tools intentionally require server launch from the worktree, `get_runtime_context` and preflight must fail early with a clear message before any review attempt. 3. Passing a worktree path must either: - fully redirect mutation workspace validation, or - be rejected as metadata-only with a clear error. 4. `dry_run_pr_review`, `mark_final_review_decision`, and `submit_pr_review` need an explicit workspace/worktree binding check or must inherit the validated active task workspace. 5. The error message should say exactly how to fix it: - relaunch/reconnect MCP server from the branches worktree, or - set the required environment variable, or - use the supported workspace-scoping mechanism. ## Acceptance criteria 1. Add tests proving reviewer mutations succeed when the active workspace is a clean `branches/` worktree. 2. Add tests proving mutations fail closed when the MCP server is rooted at the control checkout. 3. Add tests proving preflight and mutation tools agree on the same workspace root. 4. Add tests proving a metadata-only `worktree_path` cannot produce a misleading `preflight_ready: true` state when mutation tools will still fail. 5. Add a canonical handoff/error message for stale MCP server workspace binding. 6. Confirm root checkout remains protected and cannot be used for author/reviewer mutations. ## Non-goals - Do not weaken #274/#475 branches-only mutation guards. - Do not allow reviewer mutations from the stable control checkout. - Do not bypass existing capability or lease gates. ## Related - #274 — branches-only author mutation guard - #475 — hard guard preventing root checkout drift - #460 — branches-only guard vs runtime_context worktree resolution - #487 — observed failure session (if applicable)
jcwalker3 commented 2026-07-08 02:25:58 -05:00
Author
Owner
Copy Link
Copy Source

Canonical Issue State

STATE: ready-for-author
WHO_IS_NEXT: author
NEXT_ACTION: Fix reviewer mutation workspace binding so tools use the active branches worktree instead of stale MCP process root
NEXT_PROMPT:

You are the author LLM fixing MCP workspace binding for reviewer mutations in Scaled-Tech-Consulting/Gitea-Tools (issue #503).

Goal: Reviewer mutation tools (lease acquire, dry_run_pr_review, mark_final_review_decision, submit_pr_review) must resolve workspace from the active task worktree, not only MCP server process root.

Observed failure: clean branches/review-pr487-submit worktree passes preflight with worktree_path, but mutations still fail because execution resolves to control checkout.

Hard rules:
1. Live identity/profile/runtime checks first.
2. Work only in a valid branches/ worktree.
3. Do not weaken #274/#475 branches-only guards.
4. Do not allow mutations from stable control checkout.
5. Do not self-review or merge.

Implementation:
- Unify workspace resolution across get_runtime_context, preflight, and mutation tools
- Either fully redirect mutation validation via worktree_path/env, or fail early with clear metadata-only error
- Add explicit binding check to dry_run_pr_review, mark_final_review_decision, submit_pr_review
- Error messages must say how to fix: relaunch MCP from worktree, set env var, or use supported scoping mechanism
- Add canonical handoff/error message for stale MCP server workspace binding

Tests:
- reviewer mutations succeed with clean branches/ worktree as active workspace
- mutations fail closed when MCP rooted at control checkout
- preflight and mutation tools agree on workspace root
- metadata-only worktree_path cannot yield misleading preflight_ready:true when mutations will fail

Run focused workspace-binding tests then broader MCP server tests.

WHAT_HAPPENED: Durable issue filed for reviewer MCP workspace-binding defect observed in PR #487 session
WHY: Valid clean review worktrees cannot submit formal reviews when MCP server launches from control checkout
RELATED_PRS: PR #487
BLOCKERS: none
VALIDATION: not yet implemented
LAST_UPDATED_BY: prgs-author / jcwalker3

## Canonical Issue State STATE: ready-for-author WHO_IS_NEXT: author NEXT_ACTION: Fix reviewer mutation workspace binding so tools use the active branches worktree instead of stale MCP process root NEXT_PROMPT: ```text You are the author LLM fixing MCP workspace binding for reviewer mutations in Scaled-Tech-Consulting/Gitea-Tools (issue #503). Goal: Reviewer mutation tools (lease acquire, dry_run_pr_review, mark_final_review_decision, submit_pr_review) must resolve workspace from the active task worktree, not only MCP server process root. Observed failure: clean branches/review-pr487-submit worktree passes preflight with worktree_path, but mutations still fail because execution resolves to control checkout. Hard rules: 1. Live identity/profile/runtime checks first. 2. Work only in a valid branches/ worktree. 3. Do not weaken #274/#475 branches-only guards. 4. Do not allow mutations from stable control checkout. 5. Do not self-review or merge. Implementation: - Unify workspace resolution across get_runtime_context, preflight, and mutation tools - Either fully redirect mutation validation via worktree_path/env, or fail early with clear metadata-only error - Add explicit binding check to dry_run_pr_review, mark_final_review_decision, submit_pr_review - Error messages must say how to fix: relaunch MCP from worktree, set env var, or use supported scoping mechanism - Add canonical handoff/error message for stale MCP server workspace binding Tests: - reviewer mutations succeed with clean branches/ worktree as active workspace - mutations fail closed when MCP rooted at control checkout - preflight and mutation tools agree on workspace root - metadata-only worktree_path cannot yield misleading preflight_ready:true when mutations will fail Run focused workspace-binding tests then broader MCP server tests. ``` WHAT_HAPPENED: Durable issue filed for reviewer MCP workspace-binding defect observed in PR #487 session WHY: Valid clean review worktrees cannot submit formal reviews when MCP server launches from control checkout RELATED_PRS: PR #487 BLOCKERS: none VALIDATION: not yet implemented LAST_UPDATED_BY: prgs-author / jcwalker3
jcwalker3 commented 2026-07-08 10:01:53 -05:00
Author
Owner
Copy Link
Copy Source

Status: Reconciled / terminal

Summary

  • Issue #503 is already closed.
  • PR #508 is closed unmerged as superseded.
  • PR #512 merged and closed Issue #510.
  • Issue #510 implemented namespace-aware workspace binding through namespace_workspace_binding.py and verify_preflight_purity → _resolve_namespace_mutation_context.
  • #510 / #512 subsume #503 AC1–AC6.
  • Merging #508 would reintroduce duplicate narrower reviewer-only logic, so leaving #508 closed unmerged is correct.
  • No action is needed on #503.

Mutation ledger

  • This comment is documentation-only.
  • No issue close/reopen was performed.
  • No PR merge/close/reopen was performed.
  • PR #493 was untouched.

Next state

  • #503 remains closed.
  • #508 remains closed unmerged as superseded.
  • #510 remains closed.
  • #512 remains merged.
  • Resume #493 separately only after a fresh live refetch confirms it is still open and still conflicted.
## Status: Reconciled / terminal ### Summary - Issue **#503** is already **closed**. - **PR #508** is **closed unmerged** as superseded. - **PR #512** merged and closed **Issue #510**. - Issue **#510** implemented namespace-aware workspace binding through `namespace_workspace_binding.py` and `verify_preflight_purity` → `_resolve_namespace_mutation_context`. - **#510 / #512** subsume **#503** AC1–AC6. - Merging **#508** would reintroduce duplicate narrower reviewer-only logic, so leaving **#508** closed unmerged is correct. - **No action is needed on #503.** ### Mutation ledger - This comment is **documentation-only**. - No issue close/reopen was performed. - No PR merge/close/reopen was performed. - **PR #493** was untouched. ### Next state - **#503** remains closed. - **#508** remains closed unmerged as superseded. - **#510** remains closed. - **#512** remains merged. - Resume **#493** separately only after a fresh live refetch confirms it is still open and still conflicted.
Sign in to join this conversation.
No Branch/Tag Specified
Branches Tags
Labels
Clear labels
allocator
Controller-owned work allocator
 anti-stomp
Prevent concurrent LLM session stomping
 architecture
Architecture / structural design
 bug
chore
codex
OpenAI Codex client / workflow session surface
 concurrency
Concurrent session safety
 contamination
Workflow or session contamination incident
 control-plane
MCP control-plane coordination and allocation authority
 dashboard
MCP operational dashboard/queue view
 database
Internal coordination storage (SQLite/Postgres)
 design
Design / investigation, no implementation
 documentation
Docs / runbooks
 enhancement
New feature or improvement
 gitea
Gitea MCP workflow
 glitchtip
GlitchTip integration
 important
incident
Operational or process incident requiring durable audit trail
 incident-bridge
Sentry-to-Gitea incident bridging
 integration
Integration testing
 jenkins
Jenkins integration
 labels
Label taxonomy management
 leases
Lease adopt/release/expire lifecycle
 mcp
MCP server / tooling
 mcp-health
MCP namespace and runtime health
 mcp-menu
MCP menu surface
 multi-project
Work spanning multiple monitoring projects or Gitea repos
 mutating
Mutating action; requires gating
 nice-to-have
observability
Observability, metrics, traces, error reporting
 portability
Cross-platform / portability
 preflight
Shared preflight gates before mutation
 protected-branch
Protected branch / stable-branch policy concern
 queue
Work queue visibility and allocation
 read-only
Read-only, no mutation
 reconnect
MCP client reconnect/reload recovery path
 recovery
Recovery paths for stale/foreign leases
 refactor
Code refactor / restructure
 release
Release / versioning
 reliability
Reliability / failure handling
 resumable-review
Persist and resume prepared review verdicts across sessions
 reviewer
Reviewer workflow tooling
 roadmap
Roadmap / umbrella issue
 safety
Safety rails and fail-closed mutation guards
 security
Security / trust boundary
 self-hosted
Self-hosted infrastructure integration
 sentry
Sentry error monitoring integration
 stale-runtime
Stale backend daemon / runtime-vs-master parity failures
 status:blocked
Issue is blocked
 status:in-progress
Issue is being worked on
 status:pr-open
Issue has an open pull request
 status:ready
Issue is ready for work
 terminal-lock
Terminal review lock (#332) path
 testing
Tests / test coverage
 tracker
Issue tracker hygiene / meta
 type:bug
Bug or defect
 type:feature
Feature or enhancement
 type:feature
Feature or enhancement
 type:guardrail
Safety gate or guardrail
 visibility
Workflow state visibility for LLMs/operators
 workflow
Cross-tool workflow
 workflow-hardening
LLM workflow coordination hardening
 workflow-hardening
LLM workflow coordination hardening
No labels
Milestone
No items
No Milestone
Projects
Clear projects
No projects
Assignees
Clear assignees
jcwalker3 sysadmin
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: Scaled-Tech-Consulting/Gitea-Tools#503