Repository navigation
Runaway PR candidate(s) detected 2026-11-08 #2474
Description
Runaway PR Detection — 2026-11-08
Repo: combsmith/.github-private | Open PRs scanned: 17 | Candidates: 17
Runaway PR Candidates
Open PRs exceeding a soft runaway threshold (commits >50, comments >200, automated cycles >25, open >48h with agent churn, or merge-stacked: >50 commits with <20 files — a rebase/squash candidate). Detection only — no PR is mutated. See #948 / #949 / the #860 post-mortem.
| PR | Title | Triggering metric(s) |
|---|---|---|
| #2450 | feat: implement issue #2449 — SonarCloud: miscellaneous findings | open 110h with agent churn (12 cycles, >48h) |
| #2448 | fix: treat Claude quota refusals as non-fatal in PR review health report | open 116h with agent churn (1 cycles, >48h) |
| #2368 | docs: refresh org READMEs (automated) | open 476h with agent churn (6 cycles, >48h) |
| #2366 | chore: sync 3 org-standard workflow stub(s) from combsmith/.github | open 477h with agent churn (2 cycles, >48h) |
| #2328 | feat: implement issue #2327 — [ADR-0007] Record the markets pilot's measured baseline and post-collapse cost in the package (§1, §7, §11) | open 610h with agent churn (2 cycles, >48h) |
| #2319 | Potential fix for code scanning alert no. 40: Checkout of untrusted code in a non-privileged context | open 659h with agent churn (1 cycles, >48h) |
| #2318 | Potential fix for code scanning alert no. 41: Checkout of untrusted code in a non-privileged context | open 659h with agent churn (1 cycles, >48h) |
| #2314 | docs(fleet-pause): a pause set by hand is three ordered calls, a wait and a check; a live check replaces the drill | open 659h with agent churn (2 cycles, >48h) |
| #2308 | feat: implement issue #2235 — CI Failure Analyst never runs for failed GitHub Actions jobs: the check_run trigger only fires for third-party checks | open 665h with agent churn (7 cycles, >48h) |
| #2290 | feat: implement issue #2247 — Stale-disposition retry cap does not count a fix-reviews pass that dies before posting any marker (follow-up to the edited-comment deadlock fix) | open 669h with agent churn (9 cycles, >48h) |
| #2260 | feat: implement issue #2232 — Sync-stub guard for rebase passes: after the rebase, restore a synced stub left in a third state (check and restore, no push blocker) — the rebase half of #2212 | open 682h with agent churn (15 cycles, >48h) |
| #2218 | feat: implement issue #2216 — dev-lead: an agent session can leave base-branch content in the PR worktree, and the harness commits it (follow-up to #2211) | automated cycles: 33 (>25); open 690h with agent churn (33 cycles, >48h) |
| #2217 | feat: implement issue #2211 — dev-lead fix-bot-comment built a commit that undoes the PR's whole diff — the no-op guard caught it, but PR #2198 was held with its fix intact | open 690h with agent churn (15 cycles, >48h) |
| #2135 | feat: implement issue #1973 — [Engine config · Phase 1] Versioned config/ai-engines.json as the single source of truth for providers, model catalog and per-task chains | comments: 258 (>200); open 715h with agent churn (19 cycles, >48h) |
| #1796 | feat: implement issue #1733 — feature-ideation.yml: every run fails at 'Install Python jsonschema' — requirements file missing from the pinned tooling checkout | commits: 102 (>50); comments: 246 (>200); open 1352h with agent churn (8 cycles, >48h); merge-stacked: 102 commits / 2 files — rebase/squash candidate (>50 commits, <20 files) |
| #1770 | feat: implement issue #1766 — [bug] pr-review approves PRs with unresolved threads (violates its own decision gate 4) and double-reviews the same SHA | comments: 267 (>200); open 1363h with agent churn (10 cycles, >48h) |
| #1730 | feat: implement issue #1713 — [#1269 split 1/2] shadow_mode input + total PR-output suppression (inert, nothing dispatched) | comments: 302 (>200); open 1473h with agent churn (15 cycles, >48h) |
Activity