[regression][allocator] process_work_queue still selects vision, roadmap, and umbrella containers after #844 #854
Closed
opened 2026-07-23 16:39:51 -05:00 by jcwalker3
·
2 comments
No Branch/Tag Specified
Labels
Clear labels
allocator
anti-stomp
architecture
bug
chore
codex
concurrency
contamination
control-plane
dashboard
database
design
documentation
enhancement
gitea
glitchtip
important
incident
incident-bridge
integration
jenkins
labels
leases
mcp
mcp-health
mcp-menu
multi-project
mutating
nice-to-have
observability
portability
preflight
protected-branch
queue
read-only
reconnect
recovery
refactor
release
reliability
resumable-review
reviewer
roadmap
safety
security
self-hosted
sentry
stale-runtime
status:blocked
status:in-progress
status:pr-open
status:ready
terminal-lock
testing
tracker
type:bug
type:feature
type:feature
type:guardrail
visibility
workflow
workflow-hardening
workflow-hardening
Controller-owned work allocator
Prevent concurrent LLM session stomping
Architecture / structural design
OpenAI Codex client / workflow session surface
Concurrent session safety
Workflow or session contamination incident
MCP control-plane coordination and allocation authority
MCP operational dashboard/queue view
Internal coordination storage (SQLite/Postgres)
Design / investigation, no implementation
Docs / runbooks
New feature or improvement
Gitea MCP workflow
GlitchTip integration
Operational or process incident requiring durable audit trail
Sentry-to-Gitea incident bridging
Integration testing
Jenkins integration
Label taxonomy management
Lease adopt/release/expire lifecycle
MCP server / tooling
MCP namespace and runtime health
MCP menu surface
Work spanning multiple monitoring projects or Gitea repos
Mutating action; requires gating
Observability, metrics, traces, error reporting
Cross-platform / portability
Shared preflight gates before mutation
Protected branch / stable-branch policy concern
Work queue visibility and allocation
Read-only, no mutation
MCP client reconnect/reload recovery path
Recovery paths for stale/foreign leases
Code refactor / restructure
Release / versioning
Reliability / failure handling
Persist and resume prepared review verdicts across sessions
Reviewer workflow tooling
Roadmap / umbrella issue
Safety rails and fail-closed mutation guards
Security / trust boundary
Self-hosted infrastructure integration
Sentry error monitoring integration
Stale backend daemon / runtime-vs-master parity failures
Issue is blocked
Issue is being worked on
Issue has an open pull request
Issue is ready for work
Terminal review lock (#332) path
Tests / test coverage
Issue tracker hygiene / meta
Bug or defect
Feature or enhancement
Feature or enhancement
Safety gate or guardrail
Workflow state visibility for LLMs/operators
Cross-tool workflow
LLM workflow coordination hardening
LLM workflow coordination hardening
Milestone
No items
No Milestone
Projects
Clear projects
No projects
Notifications
Due Date
No due date set.
Dependencies
No dependencies set.
Reference: Scaled-Tech-Consulting/Gitea-Tools#854
Reference in New Issue
Block a user
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
Problem
On live
prgs/Scaled-Tech-Consulting/Gitea-Tools,gitea_allocate_next_work(allocation_mode=cross_role, apply=false)still selects non-implementable coordination containers after issue #844 was closed and its fix is present on current master.The live allocator correctly excludes #631 as
epic_or_child_only_container, but it does not apply the same semantic exclusion to equivalent product-vision, roadmap, and umbrella records.Affected runtime and identity
Scaled-Tech-Consulting/Gitea-Tools/prgsgitea-controller/prgs-controller/sysadmin/ controllergitea-author/prgs-author/jcwalker3/ authorb70d5f3efa410a7a8acdd3ddaa1778f4dd4e64e9Failed transition
Expected: preview eligibility excludes semantic coordination containers before ranking and returns a bounded independently executable candidate.
Actual live sequence:
action=implement; its body says, “Do not implement features on this issue.”The three-invalid-candidate skip ceiling was exhausted with #655 still invalid. No apply call was made.
Expected behavior
Semantic container classification must cover equivalent scope evidence independently of a single title prefix. Vision, roadmap, umbrella, epic, and child-only records must be excluded before ranking and before assignment/lease mutation. Independently executable children must remain eligible.
Authoritative evidence
gitea_whoami,mcp_get_control_plane_guide,mcp_check_workflow_skill_preflight,gitea_assess_master_parity, andgitea_resolve_task_capabilityproved repository, identity, workflow, parity, and safety.gitea_allocate_next_work(..., apply=false)produced fingerprints:ecd96c7adad38b22106ad918f34dd7296eb23219dc793d1c59cf16a2e3f6d9e1c4fb5b80a07b6aded347c12ff4d28603471a8c49ad8b824044dcc3a92fb4209c158905be098404bb1be3ed9675e14d278e919ba1e997f35cfd7da0da7df01ba66bd19597c06a2bcf2799a34ac267ff7841679e9c5c9fae0830307bcd712ddad1gitea_view_issueconfirmed the non-implementation scope in #652, #653, and #655.gitea_list_workflow_leasesshowed no lease was created for these previews.gitea_list_issues(state=all, limit=1000)returned 471 issues. Search terms included allocator, process_work_queue, epic, child-only, container, umbrella, roadmap, product vision, pre-rank, and regression. No open issue covers this post-closure regression. Closed #844 is the predecessor; closed #776 is related pre-rank infrastructure, not this live semantic regression.Recovery attempts and partial mutations
Recovery used the canonical per-run
exclude_issue_numberspreview path for exactly three invalid candidates. It advanced selection but did not reach bounded work.No assignment, workflow lease, issue lock, Git ref, branch, worktree, file edit, commit, push, PR, review, or merge mutation was produced by the failed allocation path.
Scheduled-run impact
Unattended cycles cannot safely apply the controller selection: without semantic validation they may allocate a non-implementable container; with validation they exhaust the bounded skip ceiling without reaching executable work.
Acceptance criteria
Non-goals
Related
Canonical Issue State
STATE: needs-author
WHO_IS_NEXT: author
NEXT_ACTION: Implement issue #854 acceptance criteria by extending semantic pre-rank container classification and regression coverage; validate; open one PR against master.
WHAT_HAPPENED: A live controller cycle on current master exhausted three exclusions and still selected #655, after #652 and #653 were also proven non-implementable containers. A post-closure regression defect was recorded as #854.
WHY: The live #844 implementation excludes the #631 epic shape but still ranks product-vision, roadmap, and umbrella child-only records for author implementation.
RELATED_PRS: none
BLOCKERS: process_work_queue semantic container exclusion is incomplete on current master.
VALIDATION: Live dry-run allocator previews, issue reads, 471-issue duplicate inventory, authoritative lease inspection, workflow proof, identity/capability checks, and master parity at
b70d5f3efa.LAST_UPDATED_BY: sysadmin (prgs-controller)
Canonical Handoff
Material new evidence — canonical handoff is itself allocator-ineligible
A fresh controller cycle on live
prgs/Scaled-Tech-Consulting/Gitea-Toolsat server/local/live-master SHAef14622ba0badd0cd7d654c05af13aca7d81bc05reproduced #854 and exposed an additional self-blocking failure:a58f31aeb966da15cdc2bdc0acc2d7f3c4c4995490976bce2464f72c6d56e19e; #653dfe1c9edd59560d67510b7a53417003077b7ff36d5782440bfdeb5bb8fd1eac9; #655a4c73f3cdbd364445deab2637ade776e38a9b749942c04f7e4ef316b6f422ff8.no_safe_workand classified #854 itself asepic_or_child_only_containerbecause its problem statement mentions “coordination container.” Fingerprint:f6683a038252c30d3617d9c454f971ce2ae10652c4d6e0265516dcb7444b77cd.status:ready, bounded, independently executable, has no linked open PR, conflicting branch, claim, or workflow lease, and already carries a canonicalneeds-authorhandoff.exclude_issue_numbersis negative-only. Supplying #854 throughcandidates_jsonstill invokes the faulty classifier.Required regression: an actionable allocator-defect issue that discusses container exclusion must remain eligible. Quoted/problem-description vocabulary must not be treated as authoritative scope delegation.
No apply call was made. The allocator probes created no assignment, lease, lock, Git ref, branch, worktree, file edit, commit, push, PR, review, or merge mutation.
Canonical Issue State
STATE: blocked
WHO_IS_NEXT: user
NEXT_ACTION: Repair or expose a sanctioned controller path that can allocate this specifically validated canonical handoff without treating incidental defect-description text as child-only scope; then rerun the full controller allocation and author workflow for #854.
NEXT_PROMPT:
WHAT_HAPPENED: #854 covers the invalid-candidate chain, but the allocator also false-positively excludes #854 itself, preventing its needs-author handoff from receiving an assignment and lease.
WHY: Semantic classification treats incidental “coordination container” text in the defect description as authoritative child-only scope evidence.
RELATED_PRS: none
BLOCKERS: The sanctioned allocator has no positive pin/route operation for a validated canonical handoff and false-positively excludes #854. Unblock condition: deploy such a route or repair the classifier, then prove a live dry-run selects #854 for author.
VALIDATION: Live controller identity/capability proof, workflow mount proof, parity at
ef14622ba0, live issue/comment reads, duplicate-work gate, three bounded queue previews, and focused candidates_json replay.LAST_UPDATED_BY: sysadmin (prgs-controller)
Canonical Handoff