Skip to content
Agentmart
Submit a tool

hunt-bug

Use when the user reports something broken, a failing case, a crash, a stack trace or unexpected behavior and wants the cause found or fixed ("why does X…

Skill by FMorschel

IndexedMIT
Write a review

Chalk Labs

Need help deploying hunt-bug?

Chalk Labs sets up agents like this inside real teams: connected to your tools, tested on your data, with a human in the loop where it matters.

Book a call

Install

Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.

npx agentmarket install skill/fmorschel-hunt-bug

Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.

Get the weekly diff

The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.

We use your email only for this newsletter. Privacy

About

Use when the user reports something broken, a failing case, a crash, a stack trace or unexpected behavior and wants the cause found or fixed ("why does X happen", "fix this", "this is broken"). Runs the bug-hunt loop (hypothesis former → verifier, repeated), then, only if a fix is wanted, the regression-test-first fix flow.

Use cases: testing

No reviews yet. Installed it? Reviews come from verified installers only. How reviews work

Permissions

What this skill can do and what it asks for, from its SKILL.md and code. Checked Nov 1, 2026.

Tools
No tools found
Secrets
None asked for
Install scripts
None found

Network hosts it talks to 1

  • codeload.github.com · found in install command

Files it touches

  • ~/.claude/skills/hunt-bug

What's inside

Skill name
hunt-bug
Description
Use when the user reports something broken, a failing case, a crash, a stack trace or unexpected behavior and wants the cause found or fixed ("why does X happen", "fix this", "this is broken"). Runs the bug-hunt loop (hypothesis former → verifier, repeated), then, only if a fix is wanted, the regression-test-first fix flow.
License
MIT
Instructions
29 lines in SKILL.md · view on GitHub

Files (1)

  • SKILL.md

SKILL.md (first lines)

## Hunting (and fixing) a bug

The orchestrator's *Bug hunt* entry point as a standalone flow. Do not spawn the `orchestrator` for the diagnosis; drive the two agents yourself.

### 1. Diagnose

1. Give `bug-hypothesis-former` the report, symptoms, logs, stack traces and repro steps. It returns a ranked list of falsifiable hypotheses. If the report is too thin to form any, ask the user for the missing symptom or repro; do not guess.
2. Hand hypotheses to `bug-verifier` **one per call, most likely first**. Independent hypotheses may be verified in parallel (one `Agent` call each, same message) when they touch different code paths and none needs the previous verdict; otherwise go sequentially.
3. CONFIRMED ends the hunt. REFUTED or INCONCLUSIVE goes back to `bug-hypothesis-former` with the verifier's evidence for the next batch. Stop when confirmed or the evidence runs out, and say which.

Neither agent edits code. The confirmed repro lands in the project's `specs/regression-test-backlog.md` via `bug-verifier` (format in `write-regression-test`).

### 2. Report the diagnosis

Root cause, evidence, the backlog entry, and what a fix would touch. If the user only asked "why", stop here and offer the fix.

### 3. Fix (only when the user wants it)

The fix's MVC is the narrowest change at the confirmed root cause plus its regression test. Nearby hardening or refactors are extras to ask about.

1. **Regression test first**: `test-writer` (existing code) or `tester` (a step in an active pipeline run) turns the backlog entry into a real test in the project's suite (see `write-regression-test`). Confirm it fails on the current code.
2. `implementer` makes it pass with the smallest change (`dart-edit-protocol`).
3. Run `evaluate-code`'s selection on the diff only, scoped to the fix (usually `architecture-guardian`, `convention-agent`, plus `test-adequacy-reviewer`).
4. A fix with no regression test is incomplete. Only an explicit "skip the regression test" from the user clears that, and then say so plainly and leave the backlog entry at "pending regression test". Otherwise set its Status to "regression test merged".
5. Offer `compose-commits` for the result.

If the fix turns out to need an API change or a new component, stop and ask rather than widening into the full pipeline; `plan-change` is the route if the user agrees.

From FMorschel/agents, MIT.

More from FMorschel/agents

Signals

Stars
0
Forks
0
Open issues
0
Last commit (this folder)
Oct 18, 2026 (22 days ago)

From GitHub (fetched Oct 24, 2026).

Details

License
MIT
First seen
Oct 24, 2026
Last checked
Oct 24, 2026

More from FMorschel

Done for you

Rather not set it up yourself?

Chalk Labs scopes, builds and runs agents for your team: the task audit, the eval harness, the integrations and the weekly ops report. Fixed scope, priced before we start.

See how we build AI agents →Read the case studies

Chalk Labs is a two-founder AI and growth studio. We build AI agents, CRMs and MVPs for Web3 and AI companies, and have shipped 50+ product and token launches.

launches shipped
50+
launches shipped
media placements
500+
media placements
founders on every account
2
founders on every account

Press we've placed: Forbes · CoinDesk · Decrypt · Cointelegraph · The Block

Rahil Jain, Co-founder, Chalk Labs. Ex-Goldman Sachs engineer. You talk to the person who builds it.

Get hunt-bug running in your business

Tell us what you want to automate. A founder replies within one working day with a plan, not a pitch deck.

Goes to the Chalk Labs founders, not the newsletter. Privacy

Need help deploying hunt-bug?

CallGet help
Cite this page

hunt-bug skill. Agent Mart. https://agentmart.dev/skills/fmorschel-hunt-bug (last updated 2026-10-24)