Swarm health manager log

[agent] Swarm manager: run 2026-10-30T09:00Z

State: YELLOW (human review backlog). 2 schedule changes, 1 LEGACY recommendation, 1 new proposal.

Why YELLOW: 27 Ready for review PRs wait on a human (26 without agent:needs-human), up from 19 at 05:00Z. The YELLOW threshold is 20. There are 33 open agent PRs (threshold 30). The merge queue is empty, so no PR has been approved; the last merge was Fix stray files breaking vendored Gradle (#656) #1819 at 10-29 22:47Z.
Not RED: CI is idle: 0 runs queued, 2 in progress, 0 GitHub-hosted jobs queued and 4 macOS jobs running. The merge queue is empty, the API had 4,339 left, and main's latest push run (5ca5623) is green.
Where the PRs come from: 9 of the 34 open PRs are from the unmanaged arch refactor (Sweep cli args.rs: inline, dedupe tests, trim docs #1842, Read the cargo config chain once (#1832) #1849, Sweep bun_preflight, composer_hints, context, fetch_stage #1855, Sweep hosted_unwind, lock_cli, list, commands/mod (+ rollback dead branch) #1864, One serve GET and one vendor retry loop in api/client.rs (#1869) #1870, Sweep remove.rs and repair.rs #1875, Sweep scan/discovery.rs #1884, Sweep scan/gc.rs #1892, Sweep scan/hosted.rs (helpers and tests region) #1897). 5 are ci-janitor PRs and about 18 are fixer PRs. Slowing only the fixers won't stop the backlog from growing; see "Needs a human".
Issues: 112 open. In 24 h, 39 were opened and 20 closed. 27 of the new issues still open are arch-audit and 10 are bughunt, so the arch audits drive inflow more than the bug hunts do.
Nightly: the 10-30 05:42Z full-scope run failed on test (macos-latest, 2) and test (windows-latest, 1), the second red nightly in a row. Fix Maven home tests failing on Windows nightly #1852 (Maven home on Windows), Fix Go GOWORK test failing on macOS nightly #1894 (Go GOWORK on macOS) and Fix CRLF bin e2e failing under npm 11 on macOS #1896 (npm 11 CRLF bin on macOS) target it.

Metrics

Metric	Now (10-30 09:00Z)	Prev (10-30 05:00Z)	24 h ago (10-29 09:00Z)
Runs queued / in progress	0 / 2	0 / 0	–
GH-hosted jobs queued / running	0 / 4 (macOS)	0 / 0	0 / –
ci.yml runs 4 h (red)	40 (1 full-scope dispatch)	36 (0)	–
All workflow runs 4 h (merge_group)	82 (0)	77 (0)	–
Merge queue entries (longest wait)	0	0	–
Merges 4 h / 24 h	0 / 21	0 / 23	0 / 32
Open PRs / Ready for review / approved	34 / 27 / 0	23 / 19 / 0	19 / 16 / –
Auto-merge armed	23	13	–
Oldest `Ready for review`	#1698 6.6 d; #1796 36.6 h (needs-human); #1829 18.3 h	6.4 d / 32.6 h / 14.2 h	–
Open issues / opened 24 h (bughunt) / closed 24 h	112 / 39 (10) / 20	110 / 38 (10) / 20	93 / 14 / 31
agent:needs-human issues / triaged unclaimed	19 / 93	19 / –	17 / –
API remaining at start	4339	4517	4896
depscan	not measured (403 cross-repo)	same	same

Managed routines

Routine	Cron	Enabled	Last run
Product fixer trig_01ACZyVgV7NCAnSpFAdkD2AD	`20 /2 * *` (was hourly)	yes	08:20Z run still running at 09:00Z
CI-lane fixer trig_015fMrRtXDsAPfQKCCVFUcUk (LEGACY)	`50 * * * *`	yes	OK 08:50-08:52Z
PR burn-down trig_01QvZ5vWNqFtnc5xgcXAJ9bE	`7 * * * *`	yes	OK 08:08-08:11Z
Steward trig_01Jyp8p1dUqNnWk8YYk3QCZU	`47 * * * *`	yes	OK 08:47-08:52Z
CI janitor trig_01GpWqqtkrvU1Yp12onhPXuc	`37 * * * *`	yes	08:38Z run still running (opened #1896)
Issue janitor trig_01VAr2WfvGJ97DW9KcUsrcpn	`17 1,7,13,19 * * *`	yes	OK 07:26-07:29Z
Hunt npm trig_01KDbttcCDVSiDettvdVNVyj	`45 0,6,12,18 * * *`	yes	OK 06:47-07:02Z
Hunt pip trig_018Kocx3sZfPLKer6HNLK1P5	`45 19 * * *` (was every 6 h)	yes	OK 07:46-08:04Z
Hunt Bundler trig_01Tb2QXRBVbWXX8cdzHgiuJ9	`45 2,8,14,20 * * *`	yes	08:45Z run still running
Hunt Go trig_01TuYcCKKBDiA7UqBU8rGmaX (LEGACY)	`45 3,9,15,21 * * *`	yes	OK 03:45-04:26Z
uv, Pipenv, Poetry, pnpm, Yarn Berry/classic, Bun, vlt hunts	baseline	no (HOLD)	–
Legacy burn-down trig_01CkssbwXU1Hm5Ew8XvCULeD (unmanaged)	`7 * * * *`	yes	OK 08:11-08:14Z (duplicates the managed one)

Unmanaged and enabled: arch refactor trig_016qiS4DEccutYafgpit2qUt (55 */2, 08:55Z run in progress), audit-core trig_01HsW456BQWRLq1bxEDoXeMv (OK 06:40-06:42Z), audit-cli-tests trig_01WmdhNjdQxeti1B4cuedRk2 (OK 03:40-03:43Z), CI profiler and CI perf implementer (not read this run).

Changes this run

Product fixer trig_01ACZyVgV7NCAnSpFAdkD2AD: 20 * * * * → 20 */2 * * * (applied). This is the YELLOW rule for too many PRs waiting on humans. It moves back toward hourly after 2 consecutive GREEN runs.
pip hunt trig_018Kocx3sZfPLKer6HNLK1P5: 45 1,7,13,19 * * * → 45 19 * * * (daily, applied). Issue inflow (39) exceeded outflow (20), and pip has the most open bughunt issues (5). I'm taking only this first step, because the bug hunts opened only 10 of the 39 new issues.
RECOMMENDED (LEGACY): CI-lane fixer trig_015fMrRtXDsAPfQKCCVFUcUk 50 * * * * → 50 */2 * * *. It refuses update_trigger, so a human has to apply this.

Prompt proposals

Reviewed this run: the pip hunt, the npm hunt and the CI janitor. get_session returns only metadata here, so I judged them from their GitHub output.

pip / npm hunts: 15-20 min runs, no failures. The CLI_CONTRACT.md path from P-20261028-3 is still wrong in both prompts.
CI janitor: it opened Annotate failing tests in macOS/Windows test legs #1890, Fix Go GOWORK test failing on macOS nightly #1894 and Fix CRLF bin e2e failing under npm 11 on macOS #1896 in the last ~2.5 h, for 5 open ci-janitor PRs in total. New proposal P-20261030-1.
None of the earlier open proposals has been applied; all still apply.

P-20261030-1 (open): CI janitor trig_01GpWqqtkrvU1Yp12onhPXuc: cap open ci-janitor PRs while the review backlog is high

Problem: the prompt says "There is no cap on open ci-janitor PRs: open a new one every run". The janitor now has 5 PRs waiting on human review, which adds to a backlog that already exceeds 20 Ready for review PRs. PRs that wait longer go stale and conflict with main, and then they re-run CI.
Evidence: open ci-janitor PRs: #1852 (12.8 h, Ready for review), #1854 (12.3 h, ready), #1890 (2.2 h, ready), #1894 (1.0 h) and #1896 (0.1 h, opened by session cse_011f9aeRC4w7jsAshzLQM6VL). There were 27 Ready for review PRs at 09:00Z and none merged since 10-29 22:47Z.

-- `gh pr list --repo SocketDev/socket-patch --state open --label ci-janitor --json number,title,headRefName,body,updatedAt` to see your open PRs. There is no cap on open ci-janitor PRs: open a new one every run that finds a worthwhile target. Skip any target whose files an open PR already changes.
+- `gh pr list --repo SocketDev/socket-patch --state open --label ci-janitor --json number,title,headRefName,body,updatedAt` to see your open PRs. Keep at most 3 open ci-janitor PRs. When 3 or more are open, open a new one only for a problem that blocks merging (main push or merge_group red, merge-queue evictions or stuck entries); otherwise spend step B driving your open PRs to green and `Ready for review`, and list the new target in your report. Skip any target whose files an open PR already changes.

P-20261029-1 (open): CI-lane fixer trig_015fMrRtXDsAPfQKCCVFUcUk (LEGACY): stale target list

Problem: the lane prompt hard-codes "Examples today" issue numbers, and 9 of those 10 issues are now closed. It also has a "re-verified since 2026-10-10" cutoff that is 19 days old. Every open ci-perf issue is agent:needs-human, so the hourly runs re-read stale targets and then do nothing, in 1-2 min.
Evidence: #823, #1373, #1372, #1300, #1267, #1176, #1173, #1172 and #1171 are closed; only #267 is open. The open ci-perf issues (#1720, #1670, #1512, #1248) are all agent:needs-human. No-op runs: cse_011f9aeRSMBgDtEvbURVFSBq (04:50Z) and cse_011f9aeQtDQJ6Wf4K3swSncJ (16:51-16:52Z, $0.57). Update 10-30 05:00Z: the 03:50Z run did open #1882 (agent/fix-ci-test-rng-dedup), and the 04:51Z run lasted 1 min. The stale list and the 2026-10-10 cutoff are still in the prompt. Update 10-30 09:00Z: still in the prompt; the lane opened #1891 at ~06:50Z and the 08:50Z run lasted 2 min.

-  Examples today: #823 (hermetic CLI test Command builder), #267 (Maven CI matrix gaps), the `ci-perf` issues (#1373, #1372, #1300, #1267, #1176, #1173, #1172, #1171).
+  At the start of each run, list the lane's open issues (`issues?state=open&labels=ci-perf`, then `ci-janitor`, `ci-perf-dashboard`); don't rely on a fixed example list. If every one is `agent:needs-human` and you have no open `agent/fix-ci-*` PR, exit right away.
...
-For each open issue in your lane that you haven't re-verified since 2026-10-10, check it against
+For each open issue in your lane without an `[agent]` re-verification comment in the last 7 days, check it against

P-20261028-3 (open): Bug hunts npm, pip, Bundler, Go: wrong `CLI_CONTRACT.md` path

Problem: the prompts' "Read first" lists name CLI_CONTRACT.md as if it were at the repo root, but the file is at crates/socket-patch-cli/CLI_CONTRACT.md.
Evidence: git ls-tree -r origin/main has only crates/socket-patch-cli/CLI_CONTRACT.md. npm session_011f9aeP9EDymUMSABUm8odK, pip session_011f9aeNuTVunCUboJgKUVfP. Bundler and Go have the same path.

-... plus README.md, CLI_CONTRACT.md, docs/ecosystems.md and ...
+... plus README.md, crates/socket-patch-cli/CLI_CONTRACT.md, docs/ecosystems.md and ...

(Make the same replacement in all four prompts, including Bundler's "CLI_CONTRACT.md's Gem stale-install guard section".)

P-20261028-1 (open): Issue janitor trig_01VAr2WfvGJ97DW9KcUsrcpn: `search/issues` is blocked

Problem: the Tooling section lists -X GET search/issues -f q=... as a working route, but cloud sessions get a 403 on search/*.
Evidence: 403 "sessions are bound to their configured repositories" (10-28 05:00Z); run cse_011f9aePYEavPJ8Yj1G2S3nk.

   - `.../pulls?state=all&per_page=100`
-  - `-X GET search/issues -f q=...`
+  (`search/*` returns 403 in cloud sessions. Find linked PRs from each issue's `timeline` (`cross-referenced`/`connected` events) and PR bodies in the `pulls?state=all` pages.)

Also: the prompt and name say "hourly", but the cron runs every 6 h.

P-20261027-1 (open): CI janitor trig_01GpWqqtkrvU1Yp12onhPXuc: claim before fixing PRs it doesn't own

Problem: the janitor pushes eviction and step-B fixes to other agents' PRs without an agent-status claim, so other routines redo the same work.
Evidence: #1648: janitor session_011f9aeQeQBzNvTuMrf6bWCp pushed 6d336be and b3a75fb without a claim, and the steward then claimed the same eviction.

 5. EVICTIONS. ... Recurring causes are step-B targets.
+Before pushing any fix to a PR you don't own (eviction or step-B target), claim it exactly as in step 2 (skip it if another agent holds a fresh claim), and set it back to `state: ready` when you leave. If the fix changes a user-facing behavior that another merged PR promised, add `agent:needs-human` instead.

P-20261027-2 (open): Steward trig_01Jyp8p1dUqNnWk8YYk3QCZU: re-check right before claiming

Problem: the steward claims PRs from an inventory snapshot that can be minutes old, so it redoes work another routine just pushed. It also wrote a heartbeat dated in the future.
Evidence: #1648: the claim came ~7 min after the janitor's b3a75fb push, and the heartbeat read 21:05Z on a block written at 20:57Z (session_011f9aePfVtbqAMfnVZey7aK).

-... claim a PR (`state: in-progress`, fresh heartbeat, `next: final reviewer: ...`) before you push to it, ...
+... claim a PR (`state: in-progress`, heartbeat set to the current time, `next: final reviewer: ...`) before you push to it, ... Immediately before claiming, re-fetch the PR's head SHA and comments since your inventory; if the head moved or another routine commented about the same problem, skip the PR.

Closed (last 10):

P-20261028-2 (PR burn-down): stale. Bugbot resumed: no-issues reviews on Fix Deno -g missing a one-segment path registry (#1784) #1785 (a1d89fd) and Skip on-disk gem preview in dry takeovers #1786 (16:32Z).

Not proposed (needs a human design call): the triage, steward, burn-down and janitor prompts still use GraphQL-backed commands, which return 403 in cloud sessions.

Needs a human

Review these (5 oldest ready, auto-merge armed so approval enqueues them): Fix hidden npm workspace members left unpatched (#1827) #1829 hidden npm workspace members (18.3 h), Fix hosted gem block mixing CRLF Gemfile endings (#1831) #1841 hosted gem CRLF block (16.4 h), Sweep cli args.rs: inline, dedupe tests, trim docs #1842 arch sweep of cli args.rs (16.0 h), Fix Go replace target missed in go 1.16 graphs (#1840) #1843 Go unpruned replace target (14.6 h), Read the cargo config chain once (#1832) #1849 arch cargo config chain (14.0 h). Also ready: Fix Maven home tests failing on Windows nightly #1852 (Maven home on Windows, for the red nightly), Fix pylock wiring ignoring an out-of-root -r pin (#1851) #1853, Retry pinned bundler installs in the e2e legs #1854, Sweep bun_preflight, composer_hints, context, fetch_stage #1855, Fix vendored dry-run counting in-sync patches (#1850) #1856, Fix pnpm brace and class workspace globs (#1846) #1863, Sweep hosted_unwind, lock_cli, list, commands/mod (+ rollback dead branch) #1864, Fix Poetry lock writer read three ways (#1847) #1865, Fix Cargo member globs with [..] classes (#1833) #1866, One serve GET and one vendor retry loop in api/client.rs (#1869) #1870, Fix hosted npm ignoring omit-lockfile-registry-resolved (#1871) #1872, Fix templated pip pins ignored beside pylock (#1873) #1874, Sweep remove.rs and repair.rs #1875, Fix Go module cache default read two ways (#1834) #1876, Build: share one test RNG across randomized tests (#1868) #1882, Fix Pipenv stale-install remedy built twice (#1878) #1883, Sweep scan/discovery.rs #1884, Fix silent PyPI hosted scan with no lockfile (#638) #1885, Annotate failing tests in macOS/Windows test legs #1890, Sweep scan/gc.rs #1892. Older ones: Fix Go patch delivery, audits, and reinstalls #1698 Go patch delivery (6.6 d), and Fix extras pin un-hosted beside a pylock (#1794) #1796 extras pin beside a pylock (agent:needs-human, 36.6 h).
Arch refactor throughput: trig_016qiS4DEccutYafgpit2qUt has 9 open PRs, the most of any routine, and the arch audits opened ~27 of the last 24 h's issues. I can't change it. Consider slowing the refactor to every 4 h, or adding it and the audits to the swarm manager's managed table.
Settings:
- Disable the legacy burn-down trig_01CkssbwXU1Hm5Ew8XvCULeD. It still runs at :07 beside the managed one (08:08Z and 08:11Z this hour).
- Apply the LEGACY recommendation above for the CI-lane fixer, and migrate it and the Go hunt off http_api so their schedules can be managed.
Red nightly (2 days running): review Fix Maven home tests failing on Windows nightly #1852, Fix Go GOWORK test failing on macOS nightly #1894 and Fix CRLF bin e2e failing under npm 11 on macOS #1896 first.
Swarm-manager tooling: depscan can't be read from this session (403), so the org-wide runner signal is still unmeasured. get_session returns no transcripts. The Slack send tool is missing from this run, so a RED notice couldn't be posted: either the connector doesn't offer it, or an org admin removed it (only an admin can allow it).

State

{
"hold": ["workflow:bughunt-*.yml", "trig_01C6hN4KE2LQApmQTkCiPJUB", "trig_01G1D53jGaSGG5mYXyC4XCkn", "trig_01XBjNxBCA46ptP7SYSzGTLa", "trig_01VDLwnjX8rN3yzDqqQgy4Er", "trig_01NLd29Xf3WNyeqxfH4k1VBj", "trig_01PjqRLg4UxFeSUe6D4DhvDg", "trig_01UoVQcRHgrrFotJxDWFRSj7", "trig_01VQA2sUR8brhUjzVfHyq9bE", "trig_01QRsTZhehm5Lrkqbj42wzsa", "trig_01EVLZKMQWfpejqu4fopiKQA", "trig_0156piJhgZYH8LWJ6TCHRV5B", "trig_01WH3BKQDSxpJ1YRrAGugjoV", "trig_013w4BBSfwBTgFpi5XeZoieR", "trig_016J8Uw7v3pjq7NLwqsL54NK", "trig_0191QAYmmqcUkTtH6xNXwjXN", "trig_01CZVKDdS68fCBhuqec4CDwN", "trig_01MubtjjqgwCBaaCvujLV69U", "trig_01NTdy2rEpze5NfvsG6RNs4v", "trig_01D498UYc7vvzNoCsPfj9M9x", "trig_01Pwpsh7b5GvmXmJC9xbPGCF", "trig_019jnAoKzxKirsPpNUPj16is", "workflow:*-compatibility.yml", "workflow:bench*", "workflow:*-ledger.yml(except janitor-ledger.yml, arch-audit-ledger.yml)", "workflow:installer-drift", "workflow:vlt-serve-watchdog"],
"changes": [
{"t": "2026-10-12T09:58Z", "routine": "trig_01RG8W22by47eK1tbe9jBjco", "from": "20 * * * *", "to": "20 1-23/2 * * *", "reason": "YELLOW: self-overlap (08:20Z run heartbe", "applied": false, "note": "RECOMMENDED: update_trigger re"},
{"t": "2026-10-12T11:00Z", "routine": "trig_01RG8W22by47eK1tbe9jBjco", "from": "20 * * * *", "to": "20 1-23/2 * * *", "reason": "YELLOW: self-overlap (09:20Z run heartbe", "applied": false, "note": "RECOMMENDED: update_trigger re"},
{"t": "2026-10-12T13:05Z", "routine": "trig_01RG8W22by47eK1tbe9jBjco", "from": "20 * * * *", "to": "20 1-23/2 * * *", "reason": "YELLOW: self-overlap (11:21Z run pushed ", "applied": false, "note": "RECOMMENDED: update_trigger re"},
{"t": "2026-10-12T14:05Z", "routine": "trig_01RG8W22by47eK1tbe9jBjco", "from": "20 * * * *", "to": "20 1-23/2 * * *", "reason": "YELLOW: self-overlap (12:21Z run pushed ", "applied": false, "note": "RECOMMENDED: update_trigger re"},
{"t": "2026-10-12T15:00Z", "routine": "trig_01RG8W22by47eK1tbe9jBjco", "from": "20 * * * *", "to": "20 */2 * * *", "reason": "YELLOW: 21 PRs waiting on a human (>20),", "applied": false, "note": "RECOMMENDED: update_trigger re"},
{"t": "2026-10-12T15:00Z", "routine": "trig_015fMrRtXDsAPfQKCCVFUcUk", "from": "50 * * * *", "to": "50 */2 * * *", "reason": "YELLOW: 21 PRs waiting on a human (>20);", "applied": false, "note": "RECOMMENDED: update_trigger re"},
{"t": "2026-10-12T16:05Z", "routine": "trig_01RG8W22by47eK1tbe9jBjco", "from": "20 * * * *", "to": "20 */2 * * *", "reason": "YELLOW: self-overlap, 12th time (14:21Z ", "applied": false, "note": "RECOMMENDED: update_trigger re"},
{"t": "2026-10-12T23:05Z", "routine": "trig_01RG8W22by47eK1tbe9jBjco", "from": "20 * * * *", "to": "20 */2 * * *", "reason": "YELLOW: self-overlap, 13th time (21:20Z ", "applied": false, "note": "RECOMMENDED: update_trigger re"},
{"t": "2026-10-13T00:05Z", "routine": "trig_01RG8W22by47eK1tbe9jBjco", "from": "20 * * * *", "to": "20 */2 * * *", "reason": "YELLOW: self-overlap within last 3 runs ", "applied": false, "note": "RECOMMENDED: update_trigger re"},
{"t": "2026-10-13T02:00Z", "routine": "trig_01RG8W22by47eK1tbe9jBjco", "from": "20 * * * *", "to": "20 */2 * * *", "reason": "YELLOW: self-overlap, 14th time (00:20Z ", "applied": false, "note": "RECOMMENDED: update_trigger re"},
{"t": "2026-10-13T03:00Z", "routine": "trig_01RG8W22by47eK1tbe9jBjco", "from": "20 * * * *", "to": "20 */2 * * *", "reason": "YELLOW: self-overlap, 15th time (the 01:", "applied": false, "note": "RECOMMENDED: update_trigger re"},
{"t": "2026-10-13T04:00Z", "routine": "trig_01RG8W22by47eK1tbe9jBjco", "from": "20 * * * *", "to": "20 */2 * * *", "reason": "YELLOW: self-overlap, 16th time (02:20Z ", "applied": false, "note": "RECOMMENDED: update_trigger re"},
{"t": "2026-10-13T05:00Z", "routine": "trig_01RG8W22by47eK1tbe9jBjco", "from": "20 * * * *", "to": "20 */2 * * *", "reason": "YELLOW: self-overlap, 17th time (03:20Z ", "applied": false, "note": "RECOMMENDED: update_trigger re"},
{"t": "2026-10-13T06:00Z", "routine": "trig_01RG8W22by47eK1tbe9jBjco", "from": "20 * * * *", "to": "20 */2 * * *", "reason": "YELLOW: self-overlap, 18th time (04:20Z ", "applied": false, "note": "RECOMMENDED: update_trigger re"},
{"t": "2026-10-13T07:00Z", "routine": "trig_01RG8W22by47eK1tbe9jBjco", "from": "20 * * * *", "to": "20 */2 * * *", "reason": "YELLOW: self-overlap, 19th time (05:20Z ", "applied": false, "note": "RECOMMENDED: update_trigger re"},
{"t": "2026-10-13T08:00Z", "routine": "trig_01RG8W22by47eK1tbe9jBjco", "from": "20 * * * *", "to": "20 */2 * * *", "reason": "YELLOW: self-overlap, 20th time (06:21Z ", "applied": false, "note": "RECOMMENDED: update_trigger re"},
{"t": "2026-10-13T09:00Z", "routine": "trig_01RG8W22by47eK1tbe9jBjco", "from": "20 * * * *", "to": "20 */2 * * *", "reason": "YELLOW: self-overlap, 21st time (07:20Z ", "applied": false, "note": "RECOMMENDED: update_trigger re"},
{"t": "2026-10-13T10:00Z", "routine": "trig_01RG8W22by47eK1tbe9jBjco", "from": "20 * * * *", "to": "20 */2 * * *", "reason": "YELLOW: self-overlap, 22nd time (08:20Z ", "applied": false, "note": "RECOMMENDED: update_trigger re"},
{"t": "2026-10-13T11:00Z", "routine": "trig_01RG8W22by47eK1tbe9jBjco", "from": "20 * * * *", "to": "20 */2 * * *", "reason": "YELLOW: self-overlap, 23rd time (09:20Z ", "applied": false, "note": "RECOMMENDED: update_trigger re"},
{"t": "2026-10-13T12:00Z", "routine": "trig_01RG8W22by47eK1tbe9jBjco", "from": "20 * * * *", "to": "20 */2 * * *", "reason": "YELLOW: self-overlap, 24th time (10:20Z ", "applied": false, "note": "RECOMMENDED: update_trigger re"},
{"t": "2026-10-13T12:00Z", "routine": "trig_01TB4Wh26CX826bt9q1yqKoQ", "from": "5 2,14 * * *", "to": "5 2 * * *", "reason": "YELLOW: issue inflow 15 (11 bughunt) > o", "applied": false, "note": "RECOMMENDED: not attempted, ev"},
{"t": "2026-10-13T13:00Z", "routine": "trig_01TB4Wh26CX826bt9q1yqKoQ", "from": "5 2,14 * * *", "to": "5 2 * * *", "reason": "YELLOW: issue inflow 16 (12 bughunt) > o", "applied": false, "note": "RECOMMENDED: update_trigger re"},
{"t": "2026-10-13T21:58Z", "routine": "trig_01ACZyVgV7NCAnSpFAdkD2AD", "from": "20 * * * *", "to": "20 */2 * * *", "reason": "YELLOW: product fixer overlapped itself.", "applied": true},
{"t": "2026-10-15T22:58Z", "routine": "trig_01ACZyVgV7NCAnSpFAdkD2AD", "from": "20 */2 * * *", "to": "20 */4 * * *", "reason": "YELLOW: product fixer overlapped itself.", "applied": true},
{"t": "2026-10-15T22:58Z", "routine": "trig_018Kocx3sZfPLKer6HNLK1P5", "from": "5 6,18 * * *", "to": "5 6 * * *", "reason": "YELLOW: issue inflow 19 (17 bughunt) > o", "applied": true},
{"t": "2026-10-20T19:58Z", "routine": "trig_01ACZyVgV7NCAnSpFAdkD2AD", "from": "20 */4 * * *", "to": "20 */2 * * *", "reason": "GREEN 2nd consecutive run: 0 queued jobs", "applied": true},
{"t": "2026-10-21T08:58Z", "routine": "trig_01ACZyVgV7NCAnSpFAdkD2AD", "from": "20 */2 * * *", "to": "20 */4 * * *", "reason": "YELLOW: product fixer overlapped itself.", "applied": true},
{"t": "2026-10-21T10:58Z", "routine": "trig_015fMrRtXDsAPfQKCCVFUcUk", "from": "50 * * * *", "to": "50 */2 * * *", "reason": "YELLOW: 21 PRs wait on a human (>20). Sl", "applied": false, "note": "RECOMMENDED: the legacy CI-lan"},
{"t": "2026-10-21T17:58Z", "routine": "trig_015fMrRtXDsAPfQKCCVFUcUk", "from": "50 * * * *", "to": "50 */2 * * *", "reason": "YELLOW: 22 PRs wait on a human (>20); pr", "applied": false, "note": "RECOMMENDED"},
{"t": "2026-10-22T08:58Z", "routine": "trig_015fMrRtXDsAPfQKCCVFUcUk", "from": "50 * * * *", "to": "50 */2 * * *", "reason": "YELLOW: 21 PRs wait on a human (>20), 0 ", "applied": false, "note": "RECOMMENDED"},
{"t": "2026-10-22T09:58Z", "routine": "trig_015fMrRtXDsAPfQKCCVFUcUk", "from": "50 * * * *", "to": "50 */2 * * *", "reason": "YELLOW: 21 PRs wait on a human (>20), 0 ", "applied": false, "note": "RECOMMENDED"},
{"t": "2026-10-22T10:58Z", "routine": "trig_015fMrRtXDsAPfQKCCVFUcUk", "from": "50 * * * *", "to": "50 */2 * * *", "reason": "YELLOW: 21 PRs wait on a human (>20), 0 ", "applied": false, "note": "RECOMMENDED"},
{"t": "2026-10-22T11:58Z", "routine": "trig_015fMrRtXDsAPfQKCCVFUcUk", "from": "50 * * * *", "to": "50 */2 * * *", "reason": "YELLOW: 21 PRs wait on a human (>20), 0 ", "applied": false, "note": "RECOMMENDED"},
{"t": "2026-10-22T12:58Z", "routine": "trig_015fMrRtXDsAPfQKCCVFUcUk", "from": "50 * * * *", "to": "50 */2 * * *", "reason": "YELLOW: 23 PRs wait on a human (>20), 0 ", "applied": false, "note": "RECOMMENDED"},
{"t": "2026-10-22T13:58Z", "routine": "trig_015fMrRtXDsAPfQKCCVFUcUk", "from": "50 * * * *", "to": "50 */2 * * *", "reason": "YELLOW: 23 PRs wait on a human (>20), 0 ", "applied": false, "note": "RECOMMENDED"},
{"t": "2026-10-22T14:58Z", "routine": "trig_015fMrRtXDsAPfQKCCVFUcUk", "from": "50 * * * *", "to": "50 */2 * * *", "reason": "YELLOW: 23 PRs wait on a human (>20), 0 ", "applied": false, "note": "RECOMMENDED"},
{"t": "2026-10-22T15:58Z", "routine": "trig_015fMrRtXDsAPfQKCCVFUcUk", "from": "50 * * * *", "to": "50 */2 * * *", "reason": "YELLOW: 24 PRs wait on a human (>20), 0 ", "applied": false, "note": "RECOMMENDED"},
{"t": "2026-10-22T20:00Z", "routine": "trig_01ACZyVgV7NCAnSpFAdkD2AD", "from": "20 */4 * * *", "to": "20 */2 * * *", "reason": "GREEN 2nd consecutive run: 0 GH-hosted queued, merge queue <", "applied": true},
{"t": "2026-10-23T01:00Z", "routine": "trig_01ACZyVgV7NCAnSpFAdkD2AD", "from": "20 */2 * * *", "to": "20 * * * *", "reason": "GREEN 7th consecutive run: 0 GH-hosted queued/running, merge", "applied": true},
{"t": "2026-10-23T07:58Z", "routine": "trig_01ACZyVgV7NCAnSpFAdkD2AD", "from": "20 * * * *", "to": "20 */2 * * *", "reason": "YELLOW: product fixer overlapped itself (06:22Z run pushed t", "applied": true},
{"t": "2026-10-23T12:59Z", "routine": "trig_01ACZyVgV7NCAnSpFAdkD2AD", "from": "20 */2 * * *", "to": "20 */4 * * *", "reason": "YELLOW: product fixer overlapped itself (10:20Z run session_", "applied": true},
{"t": "2026-10-23T16:01Z", "routine": "trig_01ACZyVgV7NCAnSpFAdkD2AD", "from": "20 */4 * * *", "to": "20 */2 * * *", "reason": "GREEN 2nd consecutive run: 0 runs queued/in progress, merge ", "applied": true},
{"t": "2026-10-23T19:01Z", "routine": "trig_01ACZyVgV7NCAnSpFAdkD2AD", "from": "20 */2 * * *", "to": "20 * * * *", "reason": "GREEN 5th consecutive run: 0 runs queued/in progress, merge ", "applied": true},
{"t": "2026-10-23T22:01Z", "routine": "trig_01ACZyVgV7NCAnSpFAdkD2AD", "from": "20 * * * *", "to": "20 */2 * * *", "reason": "YELLOW: product fixer overlapped itself again. 20:20Z run se", "applied": true},
{"t": "2026-10-24T20:58Z", "routine": "trig_01ACZyVgV7NCAnSpFAdkD2AD", "from": "20 */2 * * *", "to": "20 */4 * * *", "reason": "YELLOW: 15 PRs wait on a human (13 at 18:58Z/19:58Z), 0 appr", "applied": true},
{"t": "2026-10-29T13:15Z", "routine": "trig_01ACZyVgV7NCAnSpFAdkD2AD", "from": "20 * * * *", "to": "20 */2 * * *", "reason": "YELLOW: 21 Ready-for-review PRs wait on a human (20 excl. #1796), 0 merges since 10-28 22:56Z; fixer adds ~1 PR/h", "applied": true},
{"t": "2026-10-29T21:03Z", "routine": "trig_01ACZyVgV7NCAnSpFAdkD2AD", "from": "20 */2 * * *", "to": "20 * * * *", "reason": "GREEN 2nd consecutive run: 0 queued, MQ entry 19 min, 4 PRs wait on a human; 78 eligible triaged issues (>40), inflow 27 > outflow 25. Back to baseline.", "applied": true},
{"t": "2026-10-30T09:00Z", "routine": "trig_01ACZyVgV7NCAnSpFAdkD2AD", "from": "20 * * * *", "to": "20 */2 * * *", "reason": "YELLOW: 27 Ready-for-review PRs wait on a human (>20), 33 open agent PRs (>30), 0 merges since 10-29 22:47Z", "applied": true},
{"t": "2026-10-30T09:00Z", "routine": "trig_018Kocx3sZfPLKer6HNLK1P5", "from": "45 1,7,13,19 * * *", "to": "45 19 * * *", "reason": "YELLOW: issue inflow 39 > outflow 20 (24 h); pip has the most open bughunt issues (5). First step only: 27 of the new open issues are arch-audit, not bug hunts", "applied": true},
{"t": "2026-10-30T09:00Z", "routine": "trig_015fMrRtXDsAPfQKCCVFUcUk", "from": "50 * * * *", "to": "50 */2 * * *", "reason": "YELLOW (review backlog): LEGACY http_api, RECOMMENDED only", "applied": false, "note": "RECOMMENDED"}
],
"baseline": {"trig_01ACZyVgV7NCAnSpFAdkD2AD": "20 * * * *", "trig_01QvZ5vWNqFtnc5xgcXAJ9bE": "7 * * * *", "trig_01Jyp8p1dUqNnWk8YYk3QCZU": "47 * * * *", "trig_01GpWqqtkrvU1Yp12onhPXuc": "37 * * * *", "trig_01VAr2WfvGJ97DW9KcUsrcpn": "17 1,7,13,19 * * *", "trig_01KDbttcCDVSiDettvdVNVyj": "45 0,6,12,18 * * *", "trig_01VQA2sUR8brhUjzVfHyq9bE": "5 1,13 * * *", "trig_01UoVQcRHgrrFotJxDWFRSj7": "5 2,14 * * *", "trig_01PjqRLg4UxFeSUe6D4DhvDg": "5 3,15 * * *", "trig_01NLd29Xf3WNyeqxfH4k1VBj": "5 4,16 * * *", "trig_01VDLwnjX8rN3yzDqqQgy4Er": "5 5,17 * * *", "trig_018Kocx3sZfPLKer6HNLK1P5": "45 1,7,13,19 * * *", "trig_01XBjNxBCA46ptP7SYSzGTLa": "5 7,19 * * *", "trig_01G1D53jGaSGG5mYXyC4XCkn": "5 8,20 * * *", "trig_01C6hN4KE2LQApmQTkCiPJUB": "5 9,21 * * *", "trig_01Tb2QXRBVbWXX8cdzHgiuJ9": "45 2,8,14,20 * * *", "trig_01TuYcCKKBDiA7UqBU8rGmaX": "45 3,9,15,21 * * *", "trig_015fMrRtXDsAPfQKCCVFUcUk": "50 * * * *", "trig_01HsW456BQWRLq1bxEDoXeMv": "40 0,6,12,18 * * *", "trig_01WmdhNjdQxeti1B4cuedRk2": "40 3,9,15,21 * * *", "trig_016qiS4DEccutYafgpit2qUt": "55 */2 * * *"},
"metrics": [
{"t": "2026-10-26T13:15Z", "state": "YELLOW (24 non-draft PRs wait on a human", "gh_hosted_queued": 0, "oldest_queued_min": 0, "mq_entries": 0, "mq_max_wait_min": 0, "merges_24h": 0, "open_prs": 26, "prs_ready_human": 24, "oldest_ready_h": 135, "open_issues": 129, "issues_opened_24h": 0, "issues_closed_24h": 1, "needs_human": 16, "api_remaining": 4792},
{"t": "2026-10-26T14:08Z", "state": "YELLOW (25 non-draft PRs wait on a human", "gh_hosted_queued": 0, "oldest_queued_min": 0, "mq_entries": 0, "mq_max_wait_min": 0, "merges_24h": 0, "open_prs": 26, "prs_ready_human": 25, "oldest_ready_h": 135, "open_issues": 129, "issues_opened_24h": 0, "issues_closed_24h": 1, "needs_human": 16, "api_remaining": 4766},
{"t": "2026-10-26T15:03Z", "state": "YELLOW (25 non-draft PRs wait on a human", "gh_hosted_queued": 0, "oldest_queued_min": 0, "mq_entries": 0, "mq_max_wait_min": 0, "merges_24h": 0, "open_prs": 26, "prs_ready_human": 25, "oldest_ready_h": 136, "open_issues": 129, "issues_opened_24h": 0, "issues_closed_24h": 1, "needs_human": 16, "api_remaining": 4817},
{"t": "2026-10-26T16:05Z", "state": "YELLOW (#1648 waiting ~137 h > 72 h). Re", "gh_hosted_queued": 0, "oldest_queued_min": 0, "mq_entries": 4, "mq_max_wait_min": 25, "merges_24h": 5, "open_prs": 22, "prs_ready_human": 16, "oldest_ready_h": 137, "open_issues": 126, "issues_opened_24h": 0, "issues_closed_24h": 4, "needs_human": 16, "api_remaining": 4768},
{"t": "2026-10-26T17:05Z", "state": "YELLOW (only trigger left: #1648 waiting", "gh_hosted_queued": 0, "oldest_queued_min": 0, "mq_entries": 0, "mq_max_wait_min": 0, "merges_24h": 19, "open_prs": 9, "prs_ready_human": 6, "oldest_ready_h": 138, "open_issues": 111, "issues_opened_24h": 0, "issues_closed_24h": 17, "needs_human": 16, "api_remaining": 4800},
{"t": "2026-10-26T17:58Z", "state": "YELLOW (only trigger left: #1648 waiting", "gh_hosted_queued": 0, "oldest_queued_min": 0, "mq_entries": 1, "mq_max_wait_min": 7, "merges_24h": 19, "open_prs": 9, "prs_ready_human": 7, "oldest_ready_h": 139, "open_issues": 112, "issues_opened_24h": 0, "issues_closed_24h": 18, "needs_human": 16, "api_remaining": 4871},
{"t": "2026-10-26T18:58Z", "state": "YELLOW (only trigger: #1648 waiting ~140", "gh_hosted_queued": 0, "oldest_queued_min": 0, "mq_entries": 0, "mq_max_wait_min": 0, "merges_24h": 25, "open_prs": 3, "prs_ready_human": 2, "oldest_ready_h": 140, "open_issues": 108, "issues_opened_24h": 0, "issues_closed_24h": 22, "needs_human": 16, "api_remaining": 4936},
{"t": "2026-10-26T19:58Z", "state": "YELLOW (only trigger: #1648 waiting ~141", "gh_hosted_queued": 0, "oldest_queued_min": 0, "mq_entries": 0, "mq_max_wait_min": 0, "merges_24h": 25, "open_prs": 3, "prs_ready_human": 2, "oldest_ready_h": 141, "open_issues": 108, "issues_opened_24h": 0, "issues_closed_24h": 22, "needs_human": 16, "api_remaining": 4930},
{"t": "2026-10-26T21:00Z", "state": "YELLOW (only trigger: #1648 waiting ~142", "gh_hosted_queued": 0, "oldest_queued_min": 0, "mq_entries": 0, "mq_max_wait_min": 0, "merges_24h": 25, "open_prs": 5, "prs_ready_human": 3, "oldest_ready_h": 142, "open_issues": 109, "issues_opened_24h": 1, "issues_closed_24h": 22, "needs_human": 16, "api_remaining": 4921},
{"t": "2026-10-26T21:58Z", "state": "YELLOW (only trigger: #1648 waiting ~143", "gh_hosted_queued": 0, "oldest_queued_min": 0, "mq_entries": 0, "mq_max_wait_min": 0, "merges_24h": 25, "open_prs": 5, "prs_ready_human": 4, "oldest_ready_h": 143, "open_issues": 109, "issues_opened_24h": 1, "issues_closed_24h": 22, "needs_human": 16, "api_remaining": 4965},
{"t": "2026-10-26T22:58Z", "state": "YELLOW (only trigger: #1648 waiting ~144", "gh_hosted_queued": 0, "oldest_queued_min": 0, "mq_entries": 0, "mq_max_wait_min": 0, "merges_24h": 25, "open_prs": 5, "prs_ready_human": 4, "oldest_ready_h": 144, "open_issues": 109, "issues_opened_24h": 1, "issues_closed_24h": 22, "needs_human": 16, "api_remaining": 4916},
{"t": "2026-10-26T23:58Z", "state": "YELLOW (only trigger: #1648 waiting ~145", "gh_hosted_queued": 0, "oldest_queued_min": 0, "mq_entries": 0, "mq_max_wait_min": 0, "merges_24h": 25, "open_prs": 5, "prs_ready_human": 4, "oldest_ready_h": 145, "open_issues": 109, "issues_opened_24h": 1, "issues_closed_24h": 22, "needs_human": 16, "api_remaining": 4871},
{"t": "2026-10-27T00:58Z", "state": "YELLOW (only trigger: #1648 waiting ~146", "gh_hosted_queued": 0, "oldest_queued_min": 0, "mq_entries": 0, "mq_max_wait_min": 0, "merges_24h": 25, "open_prs": 6, "prs_ready_human": 4, "oldest_ready_h": 146, "open_issues": 109, "issues_opened_24h": 1, "issues_closed_24h": 22, "needs_human": 16, "api_remaining": 4884},
{"t": "2026-10-27T01:58Z", "state": "YELLOW (only trigger: #1648 waiting ~147", "gh_hosted_queued": 0, "oldest_queued_min": 0, "mq_entries": 0, "mq_max_wait_min": 0, "merges_24h": 25, "open_prs": 6, "prs_ready_human": 5, "oldest_ready_h": 147, "open_issues": 109, "issues_opened_24h": 1, "issues_closed_24h": 22, "needs_human": 16, "api_remaining": 4940},
{"t": "2026-10-27T02:58Z", "state": "YELLOW (only trigger: #1648 waiting ~148", "gh_hosted_queued": 0, "oldest_queued_min": 0, "mq_entries": 0, "mq_max_wait_min": 0, "merges_24h": 25, "open_prs": 6, "prs_ready_human": 5, "oldest_ready_h": 148, "open_issues": 109, "issues_opened_24h": 1, "issues_closed_24h": 22, "needs_human": 16, "api_remaining": 4929},
{"t": "2026-10-27T03:58Z", "state": "YELLOW (only trigger: #1648 waiting ~149", "gh_hosted_queued": 0, "oldest_queued_min": 0, "mq_entries": 0, "mq_max_wait_min": 0, "merges_24h": 25, "open_prs": 6, "prs_ready_human": 5, "oldest_ready_h": 149, "open_issues": 109, "issues_opened_24h": 1, "issues_closed_24h": 22, "needs_human": 16, "api_remaining": 4961},
{"t": "2026-10-27T04:58Z", "state": "YELLOW (only trigger: #1648 waiting ~150", "gh_hosted_queued": 0, "oldest_queued_min": 0, "mq_entries": 0, "mq_max_wait_min": 0, "merges_24h": 25, "open_prs": 7, "prs_ready_human": 5, "oldest_ready_h": 150, "open_issues": 110, "issues_opened_24h": 2, "issues_closed_24h": 21, "needs_human": 16, "api_remaining": 4958},
{"t": "2026-10-27T05:59Z", "state": "YELLOW (only trigger: #1648 waiting ~151", "gh_hosted_queued": 0, "oldest_queued_min": 0, "mq_entries": 0, "mq_max_wait_min": 0, "merges_24h": 25, "open_prs": 8, "prs_ready_human": 5, "oldest_ready_h": 151, "open_issues": 110, "issues_opened_24h": 2, "issues_closed_24h": 21, "needs_human": 16, "api_remaining": 4931},
{"t": "2026-10-27T06:58Z", "state": "YELLOW (only trigger: #1648 waiting ~152", "gh_hosted_queued": 0, "oldest_queued_min": 0, "mq_entries": 0, "mq_max_wait_min": 0, "merges_24h": 25, "open_prs": 8, "prs_ready_human": 7, "oldest_ready_h": 152, "open_issues": 110, "issues_opened_24h": 2, "issues_closed_24h": 21, "needs_human": 16, "api_remaining": 4938},
{"t": "2026-10-27T07:58Z", "state": "YELLOW (only trigger: #1648 waiting ~153", "gh_hosted_queued": 0, "oldest_queued_min": 0, "mq_entries": 0, "mq_max_wait_min": 0, "merges_24h": 25, "open_prs": 8, "prs_ready_human": 7, "oldest_ready_h": 153, "open_issues": 110, "issues_opened_24h": 2, "issues_closed_24h": 21, "needs_human": 16, "api_remaining": 4946},
{"t": "2026-10-27T08:58Z", "state": "YELLOW (only trigger: #1648 waiting ~154", "gh_hosted_queued": 0, "oldest_queued_min": 0, "mq_entries": 0, "mq_max_wait_min": 0, "merges_24h": 25, "open_prs": 10, "prs_ready_human": 7, "oldest_ready_h": 154, "open_issues": 110, "issues_opened_24h": 2, "issues_closed_24h": 21, "needs_human": 16, "api_remaining": 4981},
{"t": "2026-10-27T09:58Z", "state": "YELLOW (only trigger: #1648 waiting ~155", "gh_hosted_queued": 0, "oldest_queued_min": 0, "mq_entries": 0, "mq_max_wait_min": 0, "merges_24h": 25, "open_prs": 10, "prs_ready_human": 7, "oldest_ready_h": 155, "open_issues": 110, "issues_opened_24h": 2, "issues_closed_24h": 21, "needs_human": 16, "api_remaining": 4996},
{"t": "2026-10-27T10:58Z", "state": "YELLOW (only trigger: #1648 waiting ~156", "gh_hosted_queued": 0, "oldest_queued_min": 0, "mq_entries": 0, "mq_max_wait_min": 0, "merges_24h": 25, "open_prs": 10, "prs_ready_human": 9, "oldest_ready_h": 156, "open_issues": 110, "issues_opened_24h": 2, "issues_closed_24h": 21, "needs_human": 16, "api_remaining": 4998},
{"t": "2026-10-27T11:58Z", "state": "YELLOW (only trigger: #1648 waiting ~157", "gh_hosted_queued": 0, "oldest_queued_min": 0, "mq_entries": 0, "mq_max_wait_min": 0, "merges_24h": 25, "open_prs": 10, "prs_ready_human": 9, "oldest_ready_h": 157, "open_issues": 110, "issues_opened_24h": 2, "issues_closed_24h": 21, "needs_human": 16, "api_remaining": 4998},
{"t": "2026-10-27T13:00Z", "state": "YELLOW (only trigger: #1648 waiting ~158", "gh_hosted_queued": 0, "oldest_queued_min": 0, "mq_entries": 0, "mq_max_wait_min": 0, "merges_24h": 25, "open_prs": 12, "prs_ready_human": 9, "oldest_ready_h": 158, "open_issues": 110, "issues_opened_24h": 2, "issues_closed_24h": 21, "needs_human": 16, "api_remaining": 4660},
{"t": "2026-10-27T14:01Z", "state": "YELLOW (only trigger: #1648 waiting ~159", "gh_hosted_queued": 0, "oldest_queued_min": 0, "mq_entries": 0, "mq_max_wait_min": 0, "merges_24h": 32, "open_prs": 5, "prs_ready_human": 4, "oldest_ready_h": 159, "open_issues": 105, "issues_opened_24h": 2, "issues_closed_24h": 26, "needs_human": 16, "api_remaining": 4698},
{"t": "2026-10-27T14:58Z", "state": "YELLOW (only trigger: #1648 waiting ~160", "gh_hosted_queued": 0, "oldest_queued_min": 0, "mq_entries": 0, "mq_max_wait_min": 0, "merges_24h": 34, "open_prs": 3, "prs_ready_human": 2, "oldest_ready_h": 160, "open_issues": 104, "issues_opened_24h": 2, "issues_closed_24h": 27, "needs_human": 16, "api_remaining": 4816},
{"t": "2026-10-27T15:58Z", "state": "YELLOW (only trigger: #1648 waiting ~161", "gh_hosted_queued": 0, "oldest_queued_min": 0, "mq_entries": 0, "mq_max_wait_min": 0, "merges_24h": 29, "open_prs": 3, "prs_ready_human": 2, "oldest_ready_h": 161, "open_issues": 103, "issues_opened_24h": 2, "issues_closed_24h": 25, "needs_human": 16, "api_remaining": 4834},
{"t": "2026-10-27T16:58Z", "state": "YELLOW (only trigger: #1648 waiting ~162", "gh_hosted_queued": 0, "oldest_queued_min": 0, "mq_entries": 0, "mq_max_wait_min": 0, "merges_24h": 15, "open_prs": 4, "prs_ready_human": 2, "oldest_ready_h": 162, "open_issues": 103, "issues_opened_24h": 2, "issues_closed_24h": 11, "needs_human": 16, "api_remaining": 4817},
{"t": "2026-10-27T17:58Z", "state": "YELLOW (only trigger: #1648 waiting ~163", "gh_hosted_queued": 0, "oldest_queued_min": 0, "mq_entries": 0, "mq_max_wait_min": 0, "merges_24h": 15, "open_prs": 4, "prs_ready_human": 3, "oldest_ready_h": 163, "open_issues": 103, "issues_opened_24h": 2, "issues_closed_24h": 11, "needs_human": 16, "api_remaining": 4828},
{"t": "2026-10-27T18:58Z", "state": "YELLOW (only trigger: #1648 waiting ~164", "gh_hosted_queued": 0, "oldest_queued_min": 0, "mq_entries": 0, "mq_max_wait_min": 0, "merges_24h": 9, "open_prs": 4, "prs_ready_human": 3, "oldest_ready_h": 164, "open_issues": 103, "issues_opened_24h": 1, "issues_closed_24h": 7, "needs_human": 16, "api_remaining": 4858},
{"t": "2026-10-27T19:58Z", "state": "YELLOW (only trigger: #1648 waiting ~165", "gh_hosted_queued": 0, "oldest_queued_min": 0, "mq_entries": 1, "mq_max_wait_min": 3, "merges_24h": 9, "open_prs": 4, "prs_ready_human": 3, "oldest_ready_h": 165, "open_issues": 103, "issues_opened_24h": 2, "issues_closed_24h": 7, "needs_human": 16, "api_remaining": 4874},
{"t": "2026-10-27T21:00Z", "state": "GREEN. 0 queued, 0 GH-hosted running, MQ", "gh_hosted_queued": 0, "oldest_queued_min": 0, "mq_entries": 0, "mq_max_wait_min": 0, "merges_4h": 2, "merges_24h": 11, "open_prs": 6, "prs_ready_human": 3, "oldest_ready_h": 165, "open_issues": 104, "issues_opened_24h": 2, "issues_closed_24h": 7, "needs_human": 16, "api_remaining": 4800},
{"t": "2026-10-28T01:00Z", "state": "GREEN. 0 queued/in progress, MQ empty, 7", "gh_hosted_queued": 0, "oldest_queued_min": 0, "mq_entries": 0, "mq_max_wait_min": 0, "merges_4h": 1, "merges_24h": 12, "open_prs": 9, "prs_ready_human": 7, "oldest_ready_h": 9, "open_issues": 107, "issues_opened_24h": 5, "issues_closed_24h": 7, "needs_human": 17, "api_remaining": 4870},
{"t": "2026-10-28T05:00Z", "state": "HOLD (GREEN blocked only by 10 PRs waiti", "gh_hosted_queued": 0, "oldest_queued_min": 0, "mq_entries": 0, "mq_max_wait_min": 0, "merges_4h": 0, "merges_24h": 12, "open_prs": 18, "prs_ready_human": 10, "oldest_ready_h": 12.5, "open_issues": 109, "issues_opened_24h": 8, "issues_closed_24h": 9, "needs_human": 17, "api_remaining": 4832},
{"t": "2026-10-28T09:00Z", "state": "HOLD (GREEN blocked only by 15 PRs waiti", "gh_hosted_queued": 0, "oldest_queued_min": 0, "mq_entries": 0, "mq_max_wait_min": 0, "merges_4h": 0, "merges_24h": 12, "open_prs": 22, "prs_ready_human": 15, "oldest_ready_h": 16.5, "open_issues": 110, "issues_opened_24h": 10, "issues_closed_24h": 10, "needs_human": 17, "api_remaining": 4823},
{"t": "2026-10-28T13:00Z", "state": "HOLD (19 waiting on a human; no YELLOW t", "gh_hosted_queued": 0, "oldest_queued_min": 0, "mq_entries": 0, "mq_max_wait_min": 0, "merges_4h": 0, "merges_24h": 12, "open_prs": 26, "prs_ready_human": 19, "oldest_ready_h": 20.6, "open_issues": 111, "issues_opened_24h": 11, "issues_closed_24h": 10, "needs_human": 17, "api_remaining": 4734},
{"t": "2026-10-28T17:00Z", "state": "GREEN (first since 10-27 21:00Z). 20 mer", "gh_hosted_queued": 0, "oldest_queued_min": 0, "mq_entries": 0, "mq_max_wait_min": 0, "merges_4h": 20, "merges_24h": 23, "open_prs": 11, "prs_ready_human": 6, "oldest_ready_h": 13.2, "open_issues": 91, "issues_opened_24h": 13, "issues_closed_24h": 25, "needs_human": 17, "api_remaining": 4885},
{"t": "2026-10-28T21:00Z", "state": "GREEN (2nd consecutive). All managed rou", "gh_hosted_queued": 0, "oldest_queued_min": 0, "mq_entries": 0, "mq_max_wait_min": 0, "merges_4h": 2, "merges_24h": 23, "open_prs": 14, "prs_ready_human": 9, "oldest_ready_h": 17.2, "open_issues": 94, "issues_opened_24h": 16, "issues_closed_24h": 26, "needs_human": 17, "api_remaining": 4981},
{"t": "2026-10-29T01:00Z", "state": "GREEN (3rd consecutive). All managed rou", "gh_hosted_queued": 0, "oldest_queued_min": 0, "mq_entries": 0, "mq_max_wait_min": 0, "merges_4h": 10, "merges_24h": 32, "open_prs": 9, "prs_ready_human": 5, "oldest_ready_h": 11.5, "open_issues": 87, "issues_opened_24h": 14, "issues_closed_24h": 34, "needs_human": 17, "api_remaining": 4908},
{"t": "2026-10-29T05:00Z", "state": "GREEN (4th consecutive). All managed rou", "gh_hosted_queued": 0, "oldest_queued_min": 0, "mq_entries": 0, "mq_max_wait_min": 0, "merges_4h": 0, "merges_24h": 32, "open_prs": 13, "prs_ready_human": 9, "oldest_ready_h": 15.5, "open_issues": 91, "issues_opened_24h": 14, "issues_opened_24h_bughunt": 8, "issues_closed_24h": 32, "needs_human": 17, "api_remaining": 4916},
{"t": "2026-10-29T09:00Z", "state": "NOT GREEN, below YELLOW: 16 PRs wait on ", "gh_hosted_queued": 0, "oldest_queued_min": 0, "mq_entries": 0, "mq_max_wait_min": 0, "merges_4h": 0, "merges_24h": 32, "open_prs": 19, "prs_ready_human": 16, "oldest_ready_h": 19.5, "open_issues": 93, "issues_opened_24h": 14, "issues_opened_24h_bughunt": 8, "issues_closed_24h": 31, "needs_human": 17, "api_remaining": 4896},
{"t": "2026-10-29T13:15Z", "state": "YELLOW: 21 Ready-for-review PRs wait on ", "gh_hosted_queued": 0, "oldest_queued_min": 0, "mq_entries": 0, "mq_max_wait_min": 0, "merges_4h": 0, "merges_24h": 32, "open_prs": 24, "prs_ready_human": 21, "oldest_ready_h": 23.7, "open_issues": 96, "issues_opened_24h": 16, "issues_opened_24h_bughunt": 10, "issues_closed_24h": 21, "needs_human": 17, "api_remaining": 4847},
{"t": "2026-10-29T17:00Z", "state": "GREEN (1st after YELLOW): humans approve", "gh_hosted_queued": 0, "oldest_queued_min": 0, "mq_entries": 0, "mq_max_wait_min": 0, "merges_4h": 21, "merges_24h": 33, "open_prs": 7, "prs_ready_human": 3, "oldest_ready_h": 32.6, "open_issues": 89, "issues_opened_24h": 24, "issues_opened_24h_bughunt": 11, "issues_closed_24h": 26, "needs_human": 18, "api_remaining": 4998},
{"t": "2026-10-29T21:00Z", "state": "GREEN (2nd consecutive): 0 queued, 3 GH-", "gh_hosted_queued": 0, "oldest_queued_min": 0, "mq_entries": 1, "mq_max_wait_min": 19, "merges_4h": 1, "merges_24h": 32, "open_prs": 12, "prs_ready_human": 4, "oldest_ready_h": 24.6, "open_issues": 96, "issues_opened_24h": 27, "issues_opened_24h_bughunt": 11, "issues_closed_24h": 25, "needs_human": 18, "api_remaining": 4611},
{"t": "2026-10-30T01:00Z", "state": "NEUTRAL (not GREEN: 13 Ready-for-review waiting on a human >=10; not YELLOW: <=20). 0 queued, MQ empty, API 4607, no failures/overlap", "runs_queued": 0, "runs_in_progress": 0, "gh_hosted_queued": 0, "gh_hosted_running": 0, "oldest_queued_min": 0, "ci_runs_4h": 30, "ci_runs_4h_red": "0 (1 cancelled, superseded)", "all_runs_4h": 64, "mg_runs_4h": 4, "ci_jobs_sample_4_runs": "depot 86, gh-hosted 12 (3/run), other 55", "mq_entries": 0, "mq_max_wait_min": 0, "merges_4h": 1, "merges_24h": 23, "last_merge": "#1819 10-29 ~22:48Z (5ca5623)", "open_prs": 17, "prs_draft": 2, "prs_ready_human": 13, "prs_ready_human_excl_needs_human": 12, "prs_ready_unlabelled": "#1865", "prs_approved": 0, "prs_armed": 10, "prs_red": "#1698 (needs-human, unlabelled)", "prs_conflicting": 0, "oldest_ready_h": 28.6, "oldest_ready_excl_needs_human_h": 10.3, "prs_opened_24h": 31, "prs_merged_24h": 23, "prs_closed_unmerged_24h": 0, "open_issues": 103, "issues_opened_24h": 35, "issues_opened_24h_bughunt": 11, "issues_closed_24h": 18, "needs_human": 18, "api_remaining": 4607, "main_push": "5ca5623 success 22:48Z", "main_full": "10-29 05:41Z FAILURE (windows shards); next nightly ~05:41Z", "depscan": "not measured (cross-repo blocked)", "overlap": "none (all last runs succeeded, durations 1-36 min)"},
{"t": "2026-10-30T05:00Z", "state": "NEUTRAL (not GREEN: 19 Ready-for-review waiting on a human >=10; not YELLOW: <=20, 23 agent PRs <=30). 0 queued, MQ empty, API 4517, no failures/overlap", "runs_queued": 0, "runs_in_progress": 0, "gh_hosted_queued": 0, "gh_hosted_running": 0, "oldest_queued_min": 0, "ci_runs_4h": 36, "ci_runs_4h_red": "0 (30 sampled, all pull_request success)", "all_runs_4h": 77, "mg_runs_4h": 0, "ci_jobs_sample_3_runs": "depot 29, gh-hosted 3, other 12 per run", "mq_entries": 0, "mq_max_wait_min": 0, "merges_4h": 0, "merges_24h": 23, "last_merge": "#1819 10-29 22:47Z (5ca5623)", "open_prs": 23, "prs_draft": 0, "prs_ready_human": 19, "prs_ready_human_excl_needs_human": 18, "prs_unlabelled": "#1883,#1882,#1876,#1863", "prs_approved": 0, "prs_armed": 13, "prs_conflicting": 0, "oldest_ready_h": 153, "oldest_ready_note": "#1698 (6.4 d, labelled again); #1796 needs-human 32.6 h; #1829 14.2 h", "prs_opened_24h": 34, "prs_merged_24h": 23, "prs_closed_unmerged_24h": 0, "open_issues": 110, "issues_opened_24h": 38, "issues_opened_24h_bughunt": 10, "issues_closed_24h": 20, "needs_human": 19, "api_remaining": 4517, "main_push": "5ca5623 success 22:47Z", "main_full": "10-29 05:41Z FAILURE (windows shards); next nightly ~05:41Z", "depscan": "not measured (403 cross-repo)", "overlap": "none (all latest runs succeeded; product fixer 04:20-04:53Z)"},
{"t": "2026-10-30T09:00Z", "state": "YELLOW: 27 Ready-for-review waiting on a human (>20); 33 open agent PRs (>30). CI idle, MQ empty, API 4339", "runs_queued": 0, "runs_in_progress": 2, "gh_hosted_queued": 0, "gh_hosted_running": 4, "oldest_queued_min": 0, "ci_runs_4h": 40, "ci_runs_4h_red": "1 (workflow_dispatch full-scope on agent/fix-ci-test-env-guard)", "all_runs_4h": 82, "mg_runs_4h": 0, "mq_entries": 0, "mq_max_wait_min": 0, "merges_4h": 0, "merges_24h": 21, "last_merge": "#1819 10-29 22:47Z (5ca5623)", "open_prs": 34, "prs_draft": 2, "prs_ready_human": 27, "prs_ready_human_excl_needs_human": 26, "prs_approved": "0 (inferred: no merge_group runs, MQ empty)", "prs_armed": 23, "prs_arch_refactor_open": 9, "prs_ci_janitor_open": 5, "oldest_ready_h": 157.7, "prs_opened_24h": "~34", "open_issues": 112, "issues_opened_24h": 39, "issues_opened_24h_bughunt": 10, "issues_opened_24h_open_arch_audit": 27, "issues_closed_24h": 20, "needs_human": 19, "triaged_unclaimed": 93, "api_remaining": 4339, "main_push": "5ca5623 success 22:47Z", "main_full": "10-30 05:42Z FAILURE (test macos-latest 2, windows-latest 1); fixes #1852, #1894, #1896 open", "depscan": "not measured (403 cross-repo)", "overlap": "none; product fixer 08:20Z run still running at 09:00Z"}
],
"human_note": "2026-10-29 ~15:00Z (owner, via local session): ARCHITECTURE ROUTINES RESTARTED with a new charter (file-by-file implementation review: needless abstraction, single-caller fns, duplicated logic, mechanisms, comments, fake tests). The old arch routines trig_01MwFn5EYKHCD1FNYAdbC1xN, trig_011c8MKh4EpUVLRoY9utrxKc, trig_01YJDeWJdcV9oqp7vBEdYMAe no longer exist (404) and are removed from hold. NEW, agent-owned (meta_mcp, so update_trigger works), released at baseline: audit-core trig_01HsW456BQWRLq1bxEDoXeMv `40 0,6,12,18`, audit-cli-tests trig_01WmdhNjdQxeti1B4cuedRk2 `40 3,9,15,21`, refactor trig_016qiS4DEccutYafgpit2qUt `55 */2`. Ledger: arch-audit/ledger; discussion #1830 (replaces closed #560); arch-audit-ledger.yml workflow re-enabled. | PREVIOUS: 2026-10-28 (owner, via local session): Deno is RETIRED, not just held. Never release, enable or recommend the Deno bug hunt trig_01CZVKDdS68fCBhuqec4CDwN, and don't count open Deno issues/PRs as backlog. The fixer prompt and bughunt INSTRUCTIONS.md now exclude Deno. | PREVIOUS: 2026-10-27 ~19:50Z (owner, via local session): slow ramp back up, PR queue down to 4. RELEASED from HOLD at 6h, the new baseline (supersedes the old 12h cap for these four): npm trig_01KDbttcCDVSiDettvdVNVyj, pip trig_018Kocx3sZfPLKer6HNLK1P5, bundler trig_01Tb2QXRBVbWXX8cdzHgiuJ9, Go trig_01TuYcCKKBDiA7UqBU8rGmaX (Go is LEGACY http_api, so RECOMMEND only). Every other bug hunt stays on HOLD. Product fixer trig_01ACZyVgV7NCAnSpFAdkD2AD back to hourly `20 * * * *` (owner, 2026-10-27 ~20:00Z). Issue janitor trig_01VAr2WfvGJ97DW9KcUsrcpn daily -> every 6h (new baseline). Probe workflows bughunt-<pm>.yml stay disabled (Linux-only hunts). | PREVIOUS: 2026-10-20 (owner): swarm restarted after the 10-19 usage-limit pause. ALL bug hunts are back on HOLD (do not enable any). Released from HOLD: CI profiler trig_018Aa5s4K6eCHiZpiRx9eYQH and CI perf implementer trig_018LGcKNGskkaatPMs9vvVLY (enabled by hand; created via http_api, so you can only RECOMMEND changes for them). CI-lane fixer is the legacy trig_015fMrRtXDsAPfQKCCVFUcUk (the migrated copy was deleted); also http_api. | PREVIOUS: 2026-10-12 14:06Z (human via local session): released 11 p1 bug hunts (npm, pnpm, yarn classic/berry, bun, vlt, pip, uv, poetry, pipenv, bundler) at 12h staggered; baseline = 12h, do NOT speed them past 12h. Probe workflows (bughunt-<pm>.yml) stay HELD/disabled: hunts run Linux-only. bughunt-ledger.yml re-enabled. | 2026-10-13 14:17Z: routines migrated to agent-editable copies; use the new IDs (see migration_2026_10_13). You can now APPLY schedule changes.",
"migration_2026_10_13": {"note": "Managed routines recreated via the connector so update_trigger works; old http_api copies are DISABLED backups. old->new map:", "map": {"trig_01RG8W22by47eK1tbe9jBjco": "trig_01ACZyVgV7NCAnSpFAdkD2AD", "trig_015fMrRtXDsAPfQKCCVFUcUk": "trig_01NRYJL623QA5FEnHsLNAXuv", "trig_01CkssbwXU1Hm5Ew8XvCULeD": "trig_01QvZ5vWNqFtnc5xgcXAJ9bE", "trig_01JPjo6g2e9H4gSnmSuedHKW": "trig_01Jyp8p1dUqNnWk8YYk3QCZU", "trig_01R8ysuPrTj5isuxsRdB8orb": "trig_01GpWqqtkrvU1Yp12onhPXuc", "trig_01KNb1HxawPKDwsX66Fi1nEb": "trig_01VAr2WfvGJ97DW9KcUsrcpn", "trig_01BVGYbbHdmPkG8HvdN6tTxP": "trig_01KDbttcCDVSiDettvdVNVyj", "trig_01HTRW3hDwUf4etWvcwpAq5e": "trig_01VQA2sUR8brhUjzVfHyq9bE", "trig_01TB4Wh26CX826bt9q1yqKoQ": "trig_01UoVQcRHgrrFotJxDWFRSj7", "trig_019gzNxmuqQXxDHbC7Uoi1vQ": "trig_01PjqRLg4UxFeSUe6D4DhvDg", "trig_01Jf3eLQ7p17cgz3DvnrTo6N": "trig_01NLd29Xf3WNyeqxfH4k1VBj", "trig_018Z39m8d7ky29n1n4DJ94mB": "trig_01VDLwnjX8rN3yzDqqQgy4Er", "trig_01Y2M7erbpqfRpVYicJHBFsc": "trig_018Kocx3sZfPLKer6HNLK1P5", "trig_012s6oR7LQXyJJ1A8SbhtEq6": "trig_01XBjNxBCA46ptP7SYSzGTLa", "trig_01YN8YPNENDBtnEaxS3rpbCU": "trig_01G1D53jGaSGG5mYXyC4XCkn", "trig_0195sdwRw5i8vD6dPQRER34t": "trig_01C6hN4KE2LQApmQTkCiPJUB", "trig_01LJu5wvnNmVsNJtQBbUj2WH": "trig_01Tb2QXRBVbWXX8cdzHgiuJ9"}},
"anomalies": [{"t": "2026-10-29T09:00Z", "note": "Nightly full-scope ci.yml red 2 nights running (10-28 bun lockb, 10-29 windows test shards 1,2). Main push green, so not RED. Open fix candidates #1816 (CI-lane) and #1814 (janitor). Ready-for-review backlog 9 -> 16 in 4 h with 0 merges since 22:56Z (overnight US); if >20 next run, YELLOW: slow product fixer to 2 h."}, {"t": "2026-10-29T13:15Z", "note": "YELLOW on review backlog: 21 Ready-for-review, 18 auto-merge armed, 0 approvals/merges for 14 h. Product fixer slowed 1h->2h; CI-lane (LEGACY) 1h->2h RECOMMENDED. #1811 and #1822 both edit vendor/gem.rs (CRLF Gemfile.lock vs Gemfile line writer); they will conflict. Scheduled fires 4-23 min late (burn-down 12:30Z for :07)."}, {"t": "2026-10-29T17:00Z", "note": "Owner added 3 arch routines (audit-core, audit-cli-tests, refactor) at 15:00Z; not in the managed table, so reported only. Legacy burn-down trig_01CkssbwXU1Hm5Ew8XvCULeD still enabled (16:17Z run)."}, {"t": "2026-10-30T01:00Z", "what": "Legacy burn-down trig_01CkssbwXU1Hm5Ew8XvCULeD and managed trig_01QvZ5vWNqFtnc5xgcXAJ9bE both ran 00:09-00:12Z and each posted a 'labeled Ready for review' comment on #1863 and #1864 (00:11:21Z/00:11:54Z and 00:11:22Z/00:11:57Z)."}, {"t": "2026-10-30T05:00Z", "note": "Review backlog 4->13->19 Ready-for-review across 3 runs with 0 approvals and 0 merges since 22:47Z (overnight US). If >20 next run, YELLOW slows the product fixer to 2 h. Legacy burn-down trig_01CkssbwXU1Hm5Ew8XvCULeD still enabled (ran 04:08Z beside the managed one at 04:07Z)."}],
"proposals": {"reviewed": {"trig_01ACZyVgV7NCAnSpFAdkD2AD": "2026-10-30T05:00Z", "trig_01Jyp8p1dUqNnWk8YYk3QCZU": "2026-10-30T01:00Z", "trig_01GpWqqtkrvU1Yp12onhPXuc": "2026-10-30T09:00Z", "trig_01QvZ5vWNqFtnc5xgcXAJ9bE": "2026-10-30T01:00Z", "trig_01CkssbwXU1Hm5Ew8XvCULeD": "2026-10-30T01:00Z", "trig_015fMrRtXDsAPfQKCCVFUcUk": "2026-10-30T05:00Z", "trig_01VAr2WfvGJ97DW9KcUsrcpn": "2026-10-30T01:00Z", "trig_01KDbttcCDVSiDettvdVNVyj": "2026-10-30T09:00Z", "trig_018Kocx3sZfPLKer6HNLK1P5": "2026-10-30T09:00Z", "trig_01Tb2QXRBVbWXX8cdzHgiuJ9": "2026-10-29T21:00Z", "trig_01TuYcCKKBDiA7UqBU8rGmaX": "2026-10-29T21:00Z"}, "items": [{"id": "P-20261030-1", "status": "open", "routine": "trig_01GpWqqtkrvU1Yp12onhPXuc", "name": "CI janitor", "created": "2026-10-30T09:00Z", "problem": "Prompt says no cap on open ci-janitor PRs; 5 are open (#1852,#1854,#1890,#1894,#1896) while 27 Ready-for-review PRs wait on humans. Proposed: cap at 3 except merge-blocking targets.", "evidence": "#1852 12.8h, #1854 12.3h, #1890 2.2h, #1894 1.0h, #1896 0.1h (session cse_011f9aeRC4w7jsAshzLQM6VL); 0 merges since 10-29 22:47Z"}, {"id": "P-20261029-1", "status": "open", "routine": "trig_015fMrRtXDsAPfQKCCVFUcUk", "name": "Issue fixer: CI & build lane (LEGACY)", "created": "2026-10-29T05:00Z", "problem": "The lane prompt hard-codes 'Examples today' issue numbers, 9 of 10 of them now closed, and a 're-verified since 2026-10-10' cutoff that is 19 days old. Every open ci-perf issue is agent:needs-human, so hourly runs no-op in 1-2 min after re-reading stale targets.", "evidence": "#823, #1373, #1372, #1300, #1267, #1176, #1173, #1172, #1171 closed; only #267 open. Open ci-perf issues #1720, #1670, #1512, #1248 all carry agent:needs-human. Latest run cse_011f9aeRSMBgDtEvbURVFSBq 04:50-04:52Z; prior runs 1-2 min (log 10-28/29).", "still_valid": "2026-10-30T09:00Z: change not yet in the current prompt", "evidence_update": "10-29 09:00Z: the lane did useful work this time (opened #1816 agent/fix-ci-go-gowork-windows-tests at 07:08Z, aimed at the red nightly), so the no-op half is weaker; the stale example list and 2026-10-10 cutoff are still in the prompt. The 2 h schedule recommendation is withdrawn for now. | 10-30 05:00Z: 03:50Z run opened #1882; 04:51Z run 1 min; stale list still in prompt."}, {"id": "P-20261028-3", "status": "open", "routine": "trig_01KDbttcCDVSiDettvdVNVyj, trig_018Kocx3sZfPLKer6HNLK1P5, trig_01Tb2QXRBVbWXX8cdzHgiuJ9, trig_01TuYcCKKBDiA7UqBU8rGmaX", "name": "Bug hunts npm + pip + Bundler + Go", "created": "2026-10-28T21:00Z", "problem": "Read-first list names `CLI_CONTRACT.md` at repo root; it lives at crates/socket-patch-cli/CLI_CONTRACT.md, so each run looks for a missing file.", "evidence": "`git ls-tree origin/main` has only crates/socket-patch-cli/CLI_CONTRACT.md; prompts updated 10-27 19:50Z; sessions session_011f9aeP9EDymUMSABUm8odK (npm 18:45Z), session_011f9aeNuTVunCUboJgKUVfP (pip 19:45Z).", "still_valid": "2026-10-30T09:00Z: change not yet in the current prompt", "evidence_update": "10-29 05:00Z: the Bundler prompt (\"CLI_CONTRACT.md's Gem stale-install guard section\" and \"CLI_CONTRACT.md\") and the Go prompt (LEGACY, \"plus README.md, CLI_CONTRACT.md\") use the same bare path."}, {"id": "P-20261028-1", "status": "open", "routine": "trig_01VAr2WfvGJ97DW9KcUsrcpn", "name": "Issue & discussion janitor", "created": "2026-10-28T05:00Z", "problem": "Prompt lists `-X GET search/issues -f q=...` as a working REST route, but cloud sessions return 403 on search/* ('sessions are bound to their configured repositories'); other fleet prompts already say search/* is blocked. Runs that lean on it lose duplicate/fixed-PR lookups.", "evidence": "Swarm-manager session 2026-10-28T05:00Z: `gh api -X GET search/issues` -> HTTP 403 'This GitHub API path is not available'. Janitor run cse_011f9aePYEavPJ8Yj1G2S3nk (01:18-01:21Z) lasted only 3 min, shorter than its own 5-min snapshot wait. depscan routines' prompts already document search/* as blocked.", "evidence_update": "10-28 21:00Z: line `-X GET search/issues -f q=...` still in the current prompt (updated 10-27 19:50Z). Prompt and trigger name also say \"hourly\" while cron is every 6 h.", "still_valid": "2026-10-30T09:00Z: change not yet in the current prompt"}, {"id": "P-20261027-1", "routine": "trig_01GpWqqtkrvU1Yp12onhPXuc", "name": "CI janitor", "status": "open", "problem": "Janitor pushes eviction/step-B fixes to PRs it doesn't own without an agent-status claim, so other routines can't see the work and duplicate it.", "evidence": "#1648: janitor session_011f9aeQeQBzNvTuMrf6bWCp pushed 6d336be (20:39Z) and b3a75fb (20:50Z) with no claim; steward session_011f9aePfVtbqAMfnVZey7aK then claimed #1648 for the same eviction (~20:57Z).", "still_valid": "2026-10-30T09:00Z: change not yet in the current prompt"}, {"id": "P-20261027-2", "routine": "trig_01Jyp8p1dUqNnWk8YYk3QCZU", "name": "PR final review + merge steward", "status": "open", "problem": "Steward claims PRs from a stale inventory snapshot, so it redoes work another routine just pushed; its heartbeat was also written in the future.", "evidence": "#1648: claim 'fixing merge-queue eviction' written ~20:57Z after janitor's b3a75fb push at 20:50Z; heartbeat 21:05Z on a block last updated 20:57Z (session_011f9aePfVtbqAMfnVZey7aK).", "still_valid": "2026-10-30T09:00Z: change not yet in the current prompt"}, {"id": "P-20261028-2", "status": "stale", "routine": "trig_01QvZ5vWNqFtnc5xgcXAJ9bE", "name": "PR burn-down", "created": "2026-10-28T13:00Z", "problem": "Label sweep requires a Bugbot review of the current head; Bugbot is paused (unpaid invoice), so green PRs never get `Ready for review` and each run adds more 'paused' bot replies.", "evidence": "cursor[bot] 'Bugbot is paused - unpaid invoice' on #1778 11:18Z, #1780 11:54Z, #1779 12:23Z; #1778 green since ~11:20Z, unlabelled; fixer note on #1779 12:44Z.", "closed": "2026-10-28T17:00Z", "note": "Bugbot resumed: no-issues reviews on #1785 (a1d89fd) and #1786 (16:32Z)."}]}
}

Generated by Claude Code

[agent] Swarm manager: run 2026-10-11T13:58Z: one change RECOMMENDED, none applied.

Product fixer trig_01RG8W22by47eK1tbe9jBjco: 20 * * * * → 20 1-23/2 * * *. It overlapped itself: the 12:21Z run was active until ~13:40Z, and the 13:20Z run had already started. update_trigger refused the change because the routine was created via http_api, and agents may only edit routines they created. Apply it by hand at https://claude.ai/code/routines/trig_01RG8W22by47eK1tbe9jBjco.

Generated by Claude Code

[agent] Swarm manager: run 2026-10-11T14:58Z, YELLOW. No changes applied.

RECOMMENDED again: product fixer trig_01RG8W22by47eK1tbe9jBjco from 20 * * * * to 20 1-23/2 * * *. It overlapped itself for the second run in a row: the 13:20Z run was still active at 14:30Z after the 14:20Z run had started. update_trigger refused the edit (the routine was created via http_api), so a human needs to apply it.
8 green PRs are waiting on human review, and nothing has merged since 01:02Z.

Generated by Claude Code

[agent] Swarm manager: run 2026-10-13T21:58Z, state YELLOW (self-overlap).

Changes:

Product fixer trig_01ACZyVgV7NCAnSpFAdkD2AD: 20 * * * * → 20 */2 * * * (applied). Its 20:21Z run was still pushing to Fix uv script no-sources in a parent config (#1501) #1504 at 21:49Z, about 88 min in, while the 21:21Z run opened Fix Poetry source pin over a pre-2.1 lock (#1503) #1506 and pushed to it at 21:44Z. CI, the merge queue, main and the API budget are all GREEN. 4 PRs are waiting on a human review (CI: hermetic spawns in NuGet/PyPI suites (#823) #1498, CI: hermetic spawns in Cargo/Composer suites (#823) #1502, Fix npm alias links missed from a member (#1494) #1500, Fix Bun 1.3+ bun.lockb read as pre-1.3 (#1492) #1496). I'll step it back toward hourly after 2 consecutive GREEN runs with no overlap.

Generated by Claude Code

[agent] Swarm manager: changes at 2026-10-15T22:58Z (YELLOW)

trig_01ACZyVgV7NCAnSpFAdkD2AD (product fixer): 20 */2 * * * → 20 */4 * * *. It overlapped itself: the 20:20Z run pushed to Fix lockless nested npm projects in root scans (#1554) #1591 until 22:32Z (heartbeat 22:41Z) while the 22:21Z run opened Fix PyPI vendored wheels lost to gitignore/LFS (#1572) #1593 at 22:23Z.
trig_018Kocx3sZfPLKer6HNLK1P5 (pip bug hunt): 5 6,18 * * * → 5 6 * * *. Issue inflow (19, 17 from bug hunts) beat outflow (14) over 24 h, and pip has the most open unclaimed bughunt issues of the released hunts (8).

The steward is still disabled and the CI-lane fixer is still missing (see the issue body).

Generated by Claude Code

[agent] Swarm manager: changes at 2026-10-20T19:58Z

Product fixer trig_01ACZyVgV7NCAnSpFAdkD2AD: 20 */4 * * * -> 20 */2 * * *. This is the 2nd GREEN run in a row: 0 queued jobs, merge queue empty after 10 merges in 6 h, 5 PRs waiting on a human, 4,805 API requests left. It's one step back toward the hourly baseline.

Generated by Claude Code

[agent] Swarm manager: 08:58Z changes

Product fixer trig_01ACZyVgV7NCAnSpFAdkD2AD: 20 */2 * * * → 20 */4 * * * (YELLOW: self-overlap). The 06:21Z run was still working on Fix Bun dot-dir projects read as members (#1631) #1655 at 08:28Z after the 08:20Z run had opened draft Fix repair leaving npm vendor dirs uncommittable (#1581) #1656 at 08:23Z.

Generated by Claude Code

[agent] Swarm manager: run 2026-10-22T20:00Z — 1 change

Product fixer trig_01ACZyVgV7NCAnSpFAdkD2AD: 20 */4 * * * → 20 */2 * * *. The system was GREEN for 2 runs in a row (0 GitHub-hosted jobs queued, merge queue wait ≤ 13 min, 2 PRs waiting on a human, API 4,937, no routine failing or overlapping). This is one step back toward its hourly baseline.

Generated by Claude Code

[agent] Swarm manager: changes at 2026-10-23T01:00Z (state GREEN, 7th run in a row)

Product fixer trig_01ACZyVgV7NCAnSpFAdkD2AD: 20 */2 * * * → 20 * * * * (its baseline). Its 00:20Z run took 31 min, under the ~50-min bar I set for going back to hourly. CI is idle on GitHub-hosted runners (0 queued), the merge queue is empty and flowing (28 merges in 24 h), only 2 PRs wait on a human, and 4,968 API requests were left. If its runs get long enough to overlap again, I'll slow it back to 2 h.

Generated by Claude Code

[agent] Swarm manager: 2026-10-23T07:58Z change

Product fixer trig_01ACZyVgV7NCAnSpFAdkD2AD: 20 * * * * → 20 */2 * * *. It overlapped itself: the 06:22Z run was still pushing to Fix Poetry/PDM scans re-parsing the lock (#1473, #1624) #1684 at 07:46Z while the 07:21Z run opened Fix gem supersede on Bundler < 2.2 locks (#1688) #1690. It goes back to hourly after 2 GREEN runs in a row and a run that finishes in under ~50 min.

Generated by Claude Code

[agent] Swarm manager: 2026-10-23T12:58Z change (YELLOW)

Product fixer trig_01ACZyVgV7NCAnSpFAdkD2AD: 20 */2 * * * → 20 */4 * * *. It overlapped itself: the 10:20Z run was active until ~12:31Z, and the 12:20Z run started at 12:26Z. This is its slowest allowed step. Next fire: 16:20Z.

CI is idle, the merge queue is empty, and 15 PRs wait on a human (oldest #1648, 62 h).

Generated by Claude Code

[agent] Swarm manager: changes at 2026-10-23T16:01Z

Product fixer trig_01ACZyVgV7NCAnSpFAdkD2AD: 20 */4 * * * -> 20 */2 * * *. Why: 2nd GREEN run in a row (CI idle, merge queue empty, 4 PRs waiting on a human, API 4,999, no routine failing or overlapping). The last change was at 12:59Z, and its last run took 15 min.

Generated by Claude Code

[agent] Swarm manager: 2026-10-23T19:01Z change

Product fixer trig_01ACZyVgV7NCAnSpFAdkD2AD: 20 */2 * * * -> 20 * * * * (back to its baseline). GREEN for 5 runs in a row: 0 runs queued or in progress, merge queue empty, 4 PRs waiting on a human, 4,987 API requests left, no routine failing or overlapping. The last change was at 16:01Z, and its 18:21Z run took 37 min. If it overlaps itself again, it goes back to 2 h.

Generated by Claude Code

[agent] Swarm manager: changes at 2026-10-23T22:01Z

Product fixer trig_01ACZyVgV7NCAnSpFAdkD2AD: 20 * * * * → 20 */2 * * *. Reason: YELLOW. It overlapped itself for the fourth time today: the 20:20Z run ran until 21:53Z (93 min) and then stopped in REQUIRES_ACTION, while the 21:21Z run was already working. The 3 h cooldown from the 19:01Z change had expired. It stays at 2 h until several runs finish in under ~50 min.

Generated by Claude Code

[agent] Swarm manager: changes at 2026-10-24T20:58Z

Product fixer trig_01ACZyVgV7NCAnSpFAdkD2AD: 20 */2 * * * → 20 */4 * * *. 15 PRs wait on a human (up from 13), with 0 approvals and 0 merges since 02:54Z; fix: vendor hosted patches directly from lockfiles #1648 has waited ~94 h. Next fire 00:20Z.
Recommended, not applied: legacy CI-lane fixer trig_015fMrRtXDsAPfQKCCVFUcUk hourly → 2 h (http_api, can't be updated by me).

Generated by Claude Code

[agent] Swarm manager: run 2026-10-27T21:00Z, GREEN. No schedule changes. 2 new prompt proposals:

P-20261027-1, CI janitor trig_01GpWqqtkrvU1Yp12onhPXuc: claim a PR before pushing eviction or step-B fixes to it, and flag behavior changes with agent:needs-human. Evidence: fix: vendor hosted patches directly from lockfiles #1648, where the janitor pushed 6d336be and b3a75fb without a claim.
P-20261027-2, steward trig_01Jyp8p1dUqNnWk8YYk3QCZU: re-fetch the head and comments just before claiming, and set the heartbeat to the current time. Evidence: fix: vendor hosted patches directly from lockfiles #1648, where the steward claimed the eviction fix 7 min after the janitor had pushed it.

The diffs are in the issue body. Also: the legacy burn-down trig_01CkssbwXU1Hm5Ew8XvCULeD is enabled and duplicates the managed burn-down. Please disable it.

Generated by Claude Code

[agent] Swarm manager: run 2026-10-28T05:00Z

State: HOLD. The only thing keeping it from GREEN is the review backlog: 10 PRs wait on a human, 0 are approved and nothing has merged in ~7.7 h. CI, the queue and the API are all idle or healthy.
No schedule changes.
New proposal P-20261028-1 (issue janitor trig_01VAr2WfvGJ97DW9KcUsrcpn): remove the search/issues route from its Tooling list, because search/* returns 403 in cloud sessions. The diff is in the issue body.

Generated by Claude Code

[agent] Swarm manager: run 2026-10-28T13:00Z

No schedule changes. State is HOLD: 19 PRs wait on a human (GREEN needs fewer than 10; YELLOW needs more than 20), and everything else is healthy (0 queued jobs, merge queue empty, API 4,734).
New proposal P-20261028-2 (PR burn-down): Cursor Bugbot has been paused for an unpaid invoice since at least 11:18Z (Fix vendored Go undo missing modules.txt warning (#1777) #1778, Fix cargo home with HOME unset (#1135) #1779, Retry the Deno install in the deno test image #1780). The burn-down won't apply Ready for review without a Bugbot review, so green PRs now stall unlabelled. The real fix is restoring Bugbot billing. The proposal is a stop-gap for if that takes a while.

Generated by Claude Code

[agent] Swarm manager: run 2026-10-28T21:00Z. The system is GREEN, and I made no schedule changes.

New prompt proposal P-20261028-3: the npm and pip bug hunts point at CLI_CONTRACT.md at the repo root, but the file is at crates/socket-patch-cli/CLI_CONTRACT.md. The diff is in the issue body. P-20261028-1 (janitor search/issues) is still open: the line is still in the prompt.

Generated by Claude Code

[agent] Swarm manager: run 2026-10-29T05:00Z, GREEN, no schedule changes.

New proposal P-20261029-1 (CI-lane fixer trig_015fMrRtXDsAPfQKCCVFUcUk): replace the stale hard-coded "Examples today" issue list (9 of 10 closed) and the 2026-10-10 re-verify cutoff with a live lane listing and an early exit. Paired with the existing recommendation to run it every 2 h (50 */2 * * *), since it is LEGACY and only a human can change it.
P-20261028-3 extended: the Bundler and Go hunt prompts have the same wrong CLI_CONTRACT.md path (it lives at crates/socket-patch-cli/CLI_CONTRACT.md).
Heads-up: 9 PRs are Ready for review and 2 more are about to be labelled. There were 0 merges in the last 4 h for lack of approvals.

Generated by Claude Code

[agent] Swarm manager: run 2026-10-29T13:15Z, YELLOW (human-review backlog)

Applied: product fixer trig_01ACZyVgV7NCAnSpFAdkD2AD 20 * * * * → 20 */2 * * *. 21 PRs are labelled Ready for review (20 without Fix extras pin un-hosted beside a pylock (#1794) #1796), with no approval or merge since 10-28 22:56Z.
RECOMMENDED (LEGACY, a human must apply): CI-lane fixer trig_015fMrRtXDsAPfQKCCVFUcUk 50 * * * * → 50 */2 * * *.
Review first: Fix interrupted vendor eject unpatching npm (#1782) #1783, Fix Maven ${property} groupId/artifactId (#683) #1788, Fix yarn 1.0.x global packages being missed (#437) #1798, Fix Go patches ignoring GOFLAGS=-modfile (#1799) #1800, Fix Maven local repository lookup (#423) #1801. Fix cargo home tests failing on Windows nightly #1814 and CI: Fix Go workspace tests on Windows (#1815) #1816 fix the red nightly. Fix vendored gems on CRLF Gemfile.lock (#1808) #1811 and Fix vendored Gemfile line endings on CRLF (#1821) #1822 both edit vendor/gem.rs.

Generated by Claude Code

[agent] Swarm manager: run 2026-10-29T21:00Z, GREEN (2nd run in a row)

Change applied:

trig_01ACZyVgV7NCAnSpFAdkD2AD (product fixer): 20 */2 * * * → 20 * * * * (its baseline). This is the 2nd GREEN run in a row: no jobs queued, the one merge-queue entry has waited about 20 min, and 4 PRs wait on a human. There are 78 eligible triaged issues, and 27 issues opened vs 25 closed in the last 24 h.

No new prompt proposals.

Generated by Claude Code

[agent] Swarm manager: run 2026-10-30T09:00Z: YELLOW (27 Ready for review PRs wait on a human, 33 open agent PRs, 0 merges since 10-29 22:47Z; CI idle, API 4,339)

Changes:

Product fixer trig_01ACZyVgV7NCAnSpFAdkD2AD: 20 * * * * → 20 */2 * * * (applied)
pip hunt trig_018Kocx3sZfPLKer6HNLK1P5: 45 1,7,13,19 * * * → 45 19 * * * (applied). Issue inflow was 39 vs outflow 20 in 24 h, but 27 of the new open issues are arch-audit.
RECOMMENDED (LEGACY, needs a human): CI-lane fixer trig_015fMrRtXDsAPfQKCCVFUcUk 50 * * * * → 50 */2 * * *

New proposal: P-20261030-1: cap open ci-janitor PRs at 3, except for merge-blocking targets (5 are open now: #1852, #1854, #1890, #1894, #1896).

The unmanaged arch refactor has 9 open PRs, the largest share of the review backlog. See "Needs a human" in the body.

Generated by Claude Code

Navigation Menu

Repository navigation

Swarm health manager log #1382

State: YELLOW (human review backlog). 2 schedule changes, 1 LEGACY recommendation, 1 new proposal.

Metrics

Managed routines

Changes this run

Prompt proposals

P-20261030-1 (open): CI janitor trig_01GpWqqtkrvU1Yp12onhPXuc: cap open ci-janitor PRs while the review backlog is high

P-20261029-1 (open): CI-lane fixer trig_015fMrRtXDsAPfQKCCVFUcUk (LEGACY): stale target list

P-20261028-3 (open): Bug hunts npm, pip, Bundler, Go: wrong `CLI_CONTRACT.md` path

P-20261028-1 (open): Issue janitor trig_01VAr2WfvGJ97DW9KcUsrcpn: `search/issues` is blocked

P-20261027-1 (open): CI janitor trig_01GpWqqtkrvU1Yp12onhPXuc: claim before fixing PRs it doesn't own

P-20261027-2 (open): Steward trig_01Jyp8p1dUqNnWk8YYk3QCZU: re-check right before claiming

Needs a human

State

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

Swarm health manager log #1382

Description

State: YELLOW (human review backlog). 2 schedule changes, 1 LEGACY recommendation, 1 new proposal.

Metrics

Managed routines

Changes this run

Prompt proposals

P-20261030-1 (open): CI janitor trig_01GpWqqtkrvU1Yp12onhPXuc: cap open ci-janitor PRs while the review backlog is high

P-20261029-1 (open): CI-lane fixer trig_015fMrRtXDsAPfQKCCVFUcUk (LEGACY): stale target list

P-20261028-3 (open): Bug hunts npm, pip, Bundler, Go: wrong CLI_CONTRACT.md path

P-20261028-1 (open): Issue janitor trig_01VAr2WfvGJ97DW9KcUsrcpn: search/issues is blocked

P-20261027-1 (open): CI janitor trig_01GpWqqtkrvU1Yp12onhPXuc: claim before fixing PRs it doesn't own

P-20261027-2 (open): Steward trig_01Jyp8p1dUqNnWk8YYk3QCZU: re-check right before claiming

Needs a human

State

Activity

mikolalysenko commented on Oct 11, 2026

mikolalysenko commented on Oct 11, 2026

mikolalysenko commented on Oct 13, 2026

mikolalysenko commented on Oct 15, 2026

mikolalysenko commented on Oct 20, 2026

mikolalysenko commented on Oct 21, 2026

mikolalysenko commented on Oct 22, 2026

mikolalysenko commented on Oct 23, 2026

mikolalysenko commented on Oct 23, 2026

mikolalysenko commented on Oct 23, 2026

mikolalysenko commented on Oct 23, 2026

mikolalysenko commented on Oct 23, 2026

mikolalysenko commented on Oct 23, 2026

mikolalysenko commented on Oct 24, 2026

mikolalysenko commented on Oct 27, 2026

mikolalysenko commented on Oct 28, 2026

mikolalysenko commented on Oct 28, 2026

mikolalysenko commented on Oct 28, 2026

mikolalysenko commented on Oct 29, 2026

mikolalysenko commented on Oct 29, 2026

mikolalysenko commented on Oct 29, 2026

mikolalysenko commented on Oct 30, 2026

Metadata

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

Issue actions

P-20261028-3 (open): Bug hunts npm, pip, Bundler, Go: wrong `CLI_CONTRACT.md` path

P-20261028-1 (open): Issue janitor trig_01VAr2WfvGJ97DW9KcUsrcpn: `search/issues` is blocked