Starter kits

prod-error-sweep — the production log triager

Reviews production logs for actionable errors, traces one to its root cause, and drafts a verified fix — or stops without making changes if nothing…

Reviews production logs for actionable errors, traces one to its root cause, and drafts a verified fix — or stops without making changes if nothing actionable is present. Never copies credentials, tokens, PII, or private payloads out of the logs it reads.

Category: L · Monitoring & incident · Heartbeat: schedule (every 1–6 hours) · Level: L1 (report-only, drafts-not-opens-PR) · Cost: Medium

Source: Loop #4, "The production error sweep," from Forward Future's Loop Library, original prompt by Matthew Berman.

Quickstart

1. Read LOOP.md — the six parts, the prompt, the limits, the three stops,
   and the source's redaction constraint.
2. prod-error-sweep-state.md is your spine — already committed with an
   empty per-error register.
3. Claude Code: point a Routine or `claude -p` cron at
   .claude/skills/loop-task/SKILL.md. OpenCode: rename
   opencode.json.example → opencode.json, add the AGENTS.md line, wire
   skills/loop-task.md into a cron + wrapper script.
4. Rehearse once, by hand, before the schedule ever fires:
   "Run the prod-error-sweep skill now, exactly as the schedule would."
5. Grade the rehearsal with the loop-verifier agent — it scans the report
   for anything that looks like a leaked credential or PII, as an automatic
   FAIL check. PASS → arm the schedule. FAIL → fix the skill, never the
   report.

What's in this kit

File	Role
`LOOP.md`	the definition — six parts, prompt, limits, ownership, three stops
`prod-error-sweep-state.md`	the spine — per-error root cause, fix drafts, verification results (itself redaction-bound)
`loop-budget.md`	caps: 24 runs/day, 250k tokens/day — added by this library; the source specified none
`loop-constraints.md`	the constitution — the source's own credential/PII redaction rule, never opens a PR at L1
`loop-run-log.md`	append-only, one line per beat (also redaction-bound)
`.claude/skills/loop-task/SKILL.md`	the procedure (Claude Code)
`.claude/agents/loop-verifier.md`	the read-only reviewer — scans for leaked secrets as an automatic FAIL
`opencode.json.example` · `skills/loop-task.md`	the OpenCode twin — same skill text, cron + wrapper heartbeat

Loop Ready checklist

Provable success condition — the source's own spec: a report entry with a verified fix, or "nothing actionable" cleanly recorded
Run limit — 24 runs/day (added by this library)
Spine written first, committed empty
Report-only (L1) start — drafts a fix in a throwaway worktree, never opens a PR
Human gate placed — you read the report and open the real PR yourself, or promote the loop
One log line per beat, no silent runs
Kill switch — loop-pause-all in loop-budget.md

How this relates to the course

This kit uses Forward Future's Loop #4 nearly unmodified — its stopping condition ("if no actionable errors are present, stop without making changes") treats a quiet beat as a genuine success, which is exactly the opposite of the unprovable-stop anti-pattern this course warns against (infinite-loops.md, scenario 2): a loop that can correctly report "nothing found" isn't manufacturing progress. The source's own explicit constraint — never copy credentials, tokens, PII, or private payloads into prompts, PRs, or chat — is carried forward as this kit's most important rule, since production logs are exactly where that data lives.

Full attribution: resources/sources.md.