agentleFS
jamie8johnsonjamie8johnson/cqs/.claude/skills/archeo/SKILL.md

archeo

Skill18 starsChanged 5 months agoMITView on GitHub

Comment-archaeology sweep - find work items hiding in TODO/FIXME/for-now comments and plan-doc ledgers, judge them, file issues, queue trivial fixes. Good idle-loop activity.

The file

Files
SKILL.md

Comment Archaeology

Work items recorded only in comments, PR bodies, or plan ledgers get rediscovered by audits years later or never. This sweep finds them, judges them, and converts the real ones into issues (the searchable, labelable record) per the residuals-to-issues policy.

Finder pass

Scan for deferral language (ripgrep; comments and docs both):

rg -n 'TODO|FIXME|HACK|XXX|for now|temporar|deferred|follow-up|revisit|residual|inversion target|placeholder' --type rust -g '!target' | grep -v 'test'   # triage test hits separately, don't drop them
rg -n 'deferred|residual|open question|未|TBD' docs/plans/ docs/*.md

For scale, dispatch finder agents per area (src/store, src/search, src/cli, docs/plans) with a shared output format: file:line | quote | claimed reason.

Judge pass

For each hit, decide (nested judge agents for big batches; fable for judgment):

  • Already done — the deferral shipped but the comment survived → queue a comment scrub (trivial lane).
  • Stale premise — the referenced issue/feature is closed/deleted → scrub or rewrite to the surviving invariant.
  • Real work >30min — file an issue NOW (gh issue create via PowerShell, --body-file). Batch related nits into one umbrella issue — over-filing discredits the policy. The comment keeps only the timeless invariant; the work item moves to the issue.
  • Real work <30min — queue into a trivial-fix lane (lane-implementer agent, one branch for the batch).

Constraints

  • The provenance lint bans issue refs in comments — when rewriting a comment, state the invariant, never "see #1234".
  • Closing umbrella issues orphans their sub-phases — check before closing any umbrella surfaced here.
  • Plan-doc ledgers (docs/plans/*.md) claim completeness ("every deferral is listed") — verify the claim against the sweep results and amend the ledger when it lied.

Output

Report: N hits → already-done / stale / issued (#s) / queued-trivial, plus the scrub branch if one was opened. Land via /land.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.