Land functional follow-up A2: a read-only diagnostics endpoint that answers, for one run, where it is stuck, who is responsible, why it waits, the last stable checkpoint, known side effects and the next action.
Why: recovery semantics from batches 1-2 (reconciliation holds, trusted terminal records, honest failures) are only visible by querying tables by hand; operators need one authorized read that aggregates them.
Evidence: one implementation unit (internal/rundiag aggregator + GET /api/runs/{id}/diagnostics; owner/project authorization identical to neighboring run-read routes), per-state responsibility tests (recovery hold, HITL, confirmation, failed, completed, running), the architecture baseline edges registered as the reviewed design, and a green full make test (verify-a2-combo2.json). Independent review lands in the task card before merge.
Impact: backend pin advances to 06b0c9f; strictly read-only; no execution or recovery behavior changes. No UI, no write endpoints. A3/A4 remain out of scope.
Land functional follow-up A2: a read-only diagnostics endpoint that answers, for one run, where it is stuck, who is responsible, why it waits, the last stable checkpoint, known side effects and the next action.
Why: recovery semantics from batches 1-2 (reconciliation holds, trusted terminal records, honest failures) are only visible by querying tables by hand; operators need one authorized read that aggregates them.
Evidence: one implementation unit (internal/rundiag aggregator + GET /api/runs/{id}/diagnostics; owner/project authorization identical to neighboring run-read routes), per-state responsibility tests (recovery hold, HITL, confirmation, failed, completed, running), the architecture baseline edges registered as the reviewed design, and a green full `make test` (verify-a2-combo2.json). Independent review lands in the task card before merge.
Impact: backend pin advances to 06b0c9f; strictly read-only; no execution or recovery behavior changes. No UI, no write endpoints. A3/A4 remain out of scope.
Refs: docs/superpowers/plans/2026-10-21-agent-stack-audit-remediation-plan.md (A2)
Why: establish the authorized task workspace and its initial baseline.
Evidence: preserved precheck receipt in .task/feature-exec-recovery-diagnostics-a2.md; R0=c92443280ef8e9ed29bf5bf5250df650c01b4478.
Impact: task metadata only; the feature workspace starts at c92443280e.
Refs: .task/feature-exec-recovery-diagnostics-a2.md
Why: fix the read-only diagnostics contract (single GET endpoint,
aggregated responsibility/wait-reason/checkpoint/side-effects/next-step,
no write endpoints) before implementation.
Evidence: scoping over server routes, store recovery/HITL methods and
cmdbus audit records.
Impact: task metadata only.
Refs: docs/superpowers/plans/2026-10-21-agent-stack-audit-remediation-plan.md (A2)
Why: land the read-only run diagnostics endpoint (A2) per the approved
feature scope.
Evidence: integration transport (a2git2-run.json); unit tests green
(verify-a2b.json exit=0); combination re-run below.
Impact: one component gitlink.
Refs: feature/exec-recovery-diagnostics-a2
Why: the A2 diagnostics package reuses the cmdbus operation query and
store reads, mediated by the server; the three edges are the reviewed
design, recorded in the task card.
Evidence: lint-architecture passes (verify-a2arch.json); boundaries
and full combination re-run below.
Impact: three baseline edge lines.
Refs: feature/exec-recovery-diagnostics-a2
Why: audit trail for the diagnostics endpoint unit, the baseline edge
registration and the green full combination run.
Evidence: verify-a2b.json and verify-a2-combo2.json exit=0.
Impact: task metadata only.
Refs: feature/exec-recovery-diagnostics-a2
Why: close the task card inside its own PR so landing it needs no post-merge backfill or separate close PR.
Evidence: .task/feature-exec-recovery-diagnostics-a2.md -> .task/archive/feature-exec-recovery-diagnostics-a2.md; status 已完成（合入前闭合）; PR #704https://git-home.u1680832.nyat.app:30319/weichen/Agent-System-Dev/pulls/704.
Impact: task metadata only; the card is archived within PR #704 and lands with its merge commit.
Refs: task.sh close --self
Why: the intent-session tests were load-sensitive in CI (409 races
against the prior run settling); the hardening rides the A2 pin.
Evidence: repair receipt (a1ff-run.json) and 3x race verification
(verify-a1ff.json exit=0) on the merged tip; full make test green
(verify-a2-combo3.json exit=0).
Impact: one component gitlink.
Refs: PR #704
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.
Land functional follow-up A2: a read-only diagnostics endpoint that answers, for one run, where it is stuck, who is responsible, why it waits, the last stable checkpoint, known side effects and the next action.
Why: recovery semantics from batches 1-2 (reconciliation holds, trusted terminal records, honest failures) are only visible by querying tables by hand; operators need one authorized read that aggregates them.
Evidence: one implementation unit (internal/rundiag aggregator + GET /api/runs/{id}/diagnostics; owner/project authorization identical to neighboring run-read routes), per-state responsibility tests (recovery hold, HITL, confirmation, failed, completed, running), the architecture baseline edges registered as the reviewed design, and a green full
make test(verify-a2-combo2.json). Independent review lands in the task card before merge.Impact: backend pin advances to 06b0c9f; strictly read-only; no execution or recovery behavior changes. No UI, no write endpoints. A3/A4 remain out of scope.
Refs: docs/superpowers/plans/2026-10-21-agent-stack-audit-remediation-plan.md (A2)
View command line instructions
Checkout
From your project repository, check out a new branch and test the changes.