Activity

1. harlan-github-agent

   added

   routine:daily-checkin

   on Oct 8, 2026
2. harlan-github-agent commented on Oct 8, 2026

   

   harlan-github-agentbot

   on Oct 8, 2026 – with Harlan GitHub Agent

   ContributorAuthor

   More actions

   [ACTION NEEDED] Daily check-in: 2026-10-08

   2026-10-08T20:40:00.000Z — daily-checkin on skilld-dev/skilld.dev | 1 found | 1 new | 0 already known | 0 outside allowed scope | 0 over 5 files | 1 issues requested

   RED skilld.dev

   Nightly health RED on three standing reasons; the #581 deploy added a repository-purpose 404 refail, and the SKILLD-3W 503 storm outlived #575.

   Check	Result	Detail
   Front doors	Pass	HTTP 200
   Deploy and CI	Pass	324d0142 live; three workflows passed
   Protected health	Fail	5 failed jobs; 44 sync failures; token expiry
   Sentry	Warn	75 unresolved; 503 storm continues
   Database	Warn	19 failed jobs in window
   Workers	Warn	10 exceptions, 1 load shed

   Needs you

   1. Rotate the production GITHUB_TOKEN before ~2026-10-22; the nightly health gates on its 14-day expiry.
   2. Review the Cloudflare escalation draft for Escalate recurring D1 internal errors to Cloudflare #393; internal errors recurred through 17:25Z.
   3. Ready Unpinned skills answer a permanent 503 retry loop on skills-raw #577: SKILLD-3W reached 3,280 events through 20:09Z on the current release; the culprit moved from posthog to vercel/eve, so fix(registry): complete large repository source refresh #575 did not close it.

   Changed since yesterday

   • New: repository-purpose jobs retry 5 times against permanently 404 repos and re-enqueue daily; 4 of the 5 nightly failed-job faults. Candidate filed.
   • Recovered: fix(d1): retry reads blocked by a D1 import job #580 import-job read retry deployed; the specific message has not recurred.
   • Recovered: skill-sync identities now archived (44 named); ai-ready:cron quiet a second window.
   • Improved: sync-failed Skills 95 → 48 → 44.

   Unchanged work (6)

   • D1 internal-error escalation (Escalate recurring D1 internal errors to Cloudflare #393), draft awaiting review
   • X spend policy (X discovery projects past the accepted $40 monthly spend line #179), $7.25/mo estimate, below the $40 line
   • SKILLD-3V network-loss, 2 events, unchanged since 13:25Z
   • SKILLD-3X cold-key D1 overload, 1 event
   • Garbage-owner 404 writes, no fresh impact
   • PRs docs(work): add Reddit launch brief #560, feat(badge): show README badge only on indexable hubs, noindex kit #321, feat(indexnow): re-enable submission for the curated set #320, feat(seo): experiment C, unique value template on 10 Skills #318 open, no CI blockers

   Evidence

   Exit 1 | coverage complete | deployment 324d0142 | baseline 11.9h fresh
   Archive: ~/.local/state/daily-checkin/skilld-dev/skilld.dev/2026-10-08T20-52-48-813Z-c5e5f10f-344c-408b-ac29-e5d0d3e54bce.md
   Pulse: 44 users, 20,478 Skills (213 new/24h), 98 repo changes, 12 command copies (9 run, 3 install), 0 digests sent, $0 recorded AI cost, X $7.25/mo Post-only (124/2000 reads today); actual billing unavailable. Migrations match at 0149; no stale tasks.

   Since the #581 deploy, registry/repository-purpose throws "Repository purpose source unavailable: 404" for permanently gone upstream repos, so each exhausts five attempts (amnadtaowsoam/cerebraskills, demerzels-lab/elsamultiskillagent, duclm1x1/dive-ai, housegarofalo/claude-code-base today) and, with broken_since NULL, re-enqueues daily, driving 4 of the 5 nightly failed-job RED faults.

   Target: layers/registry/server/utils/repository-purpose-effect.ts plus server/jobs/registry/repository-purpose.ts

   Verify with: pnpm exec wrangler d1 execute skilld-db --remote --command "SELECT attempts, COUNT(*) n FROM failed_jobs WHERE job_type='registry/repository-purpose' GROUP BY attempts" --json returns only attempts=1 rows after a 404 repo passes through, and pnpm vitest run test/unit/repository-purpose-effect.test.ts passes

   Estimated to change 2 files.