agentleFS
KcAnomKcAnom/pi-fairy-tales-chains/skills/bughunt-verify/SKILL.md

bughunt-verify

Skill0 starsChanged 4 months agoMITView on GitHub

Phase 4 of 4 (final) in the bughunt chain (repro → diagnose → fix → verify). Re-runs the original repro to confirm it's resolved, runs the broader test suite, checks for regressions, and reports faithfully — naming any check that fails. Trigger when the previous phase says "run bughunt-verify next", or standalone via "bughunt-verify", "phase 4 of bughunt", "verify the fix", "confirm the bug is fixed".

The file

We split this skill into the rules it gives an agent

11 rules. About 0 are common, 3 are distinctive, and 8 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈3

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 8

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    All chain state is managed by the chain tool: authoritative JSON at .pi/fairy-tales/chains/bughunt/state.json with a human-readable …

    line 25
  2. Very specific:

    Call the chain tool with action: "status", chain: "bughunt" (a legacy .bughunt-state.md from an older version is imported automatically):

    line 32
  3. Very specific:

    If there is no run → abort: "No bughunt run found.

    line 35
  4. Very specific:

    If the run is not active with currentPhase: "verify" → abort: "Expected the run to be at phase verify but it is at <currentPhase/status>.

    line 36
  5. Very specific:

    If the fix data (data.checkpointBefore or the fix phase summary) is missing → abort: "Phase pointer looks right but the Phase 3 output is …

    line 39
  6. Very specific:

    If the tool reports the chain is locked by another session, abort.

    line 42
  7. Very specific:

    Run the exact data.reproCommand from Phase 1 (unchanged — do not substitute a different command).

    line 49
  8. Very specific:

    Run the project's full test command (not just the post-edit hook's scoped run from Phase 3) — e.g.

    line 58
  9. Distinctive:

    Call the chain tool:

    line 80
  10. Distinctive:

    After completing this phase, output exactly:

    line 94
  11. Distinctive:

    Do not continue. Stop here.

    line 100

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Bughunt: Verify

Chain version: 2 (durable state contract — see docs/STATE-CONTRACT.md)

Phase 4 of 4: bughunt-repro → bughunt-diagnose → bughunt-fix → bughunt-verify

Overview

Final phase. Proves the fix actually resolved the bug — the exact repro from Phase 1 no longer reproduces the failure — and that the broader test suite still passes, then closes the chain. Reports faithfully: if a check fails, it says so and points at the Phase 3 checkpoints for rollback, rather than declaring success.

All chain state is managed by the chain tool: authoritative JSON at .pi/fairy-tales/chains/bughunt/state.json with a human-readable state.md projection beside it. Never create or edit state files by hand, and never create lock files — the tool locks for you.

Load State

Call the chain tool with action: "status", chain: "bughunt" (a legacy .bughunt-state.md from an older version is imported automatically):

  • If there is no run → abort: "No bughunt run found. Run bughunt-repro first."
  • If the run is not active with currentPhase: "verify" → abort: "Expected the run to be at phase verify but it is at <currentPhase/status>. Run that phase's skill instead."
  • If the fix data (data.checkpointBefore or the fix phase summary) is missing → abort: "Phase pointer looks right but the Phase 3 output is missing. Treat as corrupted — restart the chain or repair via chain action 'update'."
  • If the tool reports the chain is locked by another session, abort; a dead session's lock can be cleared with action: "unlock". Never create lock files yourself.

Workflow

Step 1 — Re-run the original repro

Run the exact data.reproCommand from Phase 1 (unchanged — do not substitute a different command). Compare the result against the expected behavior recorded in data.expectedBehavior:

  • Matches expected → repro check ✓.
  • Still matches the old data.actualBehavior (or errors differently but still fails) → repro check ✗. Do not soften this into a partial pass.

Step 2 — Run the broader test suite

Run the project's full test command (not just the post-edit hook's scoped run from Phase 3) — e.g. .pi/test-command in full-suite form if it supports one, or the project's standard npm test / pytest / equivalent. This is what catches regressions the narrow post-edit hook run in Phase 3 wouldn't have exercised.

Step 3 — Check for regressions

Compare the broader suite's result to a known-good baseline if one is available (e.g. CI status on the pre-fix commit, or simply "did anything fail that wasn't failing before"). Name any newly-failing test explicitly — do not lump it in as "some tests failed."

Step 4 — Report

Produce a short verification report: the bug, the fix location, and the result of each check (✓ / ✗) — repro re-run, broader suite, regressions. If anything failed, say so plainly and point at data.checkpointBefore from Phase 3 as the rollback target (/rollback), and recommend re-running bughunt-fix (or bughunt-diagnose if the root cause looks wrong) rather than patching further in this phase.

Step 5 — Complete the phase

Call the chain tool:

action: "complete-phase", chain: "bughunt", phase: "verify",
summary: "verification report: repro re-run <✓/✗>; broader suite <✓/✗>; regressions <none found / listed>",
data: { "verificationReport": "<the verification report from Step 4>" }

The tool validates this is the current phase, marks the run complete, and rewrites the JSON + markdown projection atomically — this is the final phase, so the run's status becomes complete and its lock is released.

Handoff

After completing this phase, output exactly:

Chain complete! All 4 phases finished. Chain state updated (.pi/fairy-tales/chains/bughunt/state.json). Run bughunt-repro again to start a new session.

Do not continue. Stop here.

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.