agentleFS
dativo-iodativo-io/talon/.claude/skills/investigate-bug/SKILL.md

investigate-bug

Skill18 starsChanged 24 days agoApache-2.0View on GitHub

Reproduce and root-cause a Talon bug before deciding on a fix.

8 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

4 rules. 0 are common, 4 are distinctive, and 0 are very specific to this project.

  • 0

    Common

    Most of the 8 edited versions we read keep these.

  • 4

    Distinctive

    This author's own choices. Worth a look.

  • 0

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Check whether the bug is:

    line 13
  2. Distinctive:

    Search for variants of the same pattern.

    line 22
  3. Distinctive:

    Do not edit code unless the user asked for a fix as well as investigation.

    line 25
  4. Distinctive:

    Do not mutate GitHub unless explicitly authorized.

    line 26

“Common” is measured: kept by at least 6 in 10 of the 8 edited versions of this file we could read. “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

Investigate $ARGUMENTS.

  1. Establish the expected contract from current code/tests/docs/issues.
  2. Reproduce the reported behavior with the smallest reliable command or test. For a CI failure, first establish which commit CI ran and run that step's exact command locally (/verify-change has the job-to-command map).
  3. Minimize the reproducer and identify the first incorrect boundary/state transition.
  4. Trace data/identity/state from entry point to the failing effect.
  5. Check whether the bug is:
    • local implementation error;
    • duplicated semantic path;
    • stale compatibility behavior;
    • projection/classification drift;
    • persistence/migration issue;
    • race/recovery failure;
    • trust-boundary validation issue.
  6. Identify who can trigger it and the realistic blast radius.
  7. Search for variants of the same pattern.
  8. Propose the smallest fix and the regression test that proves it.

Do not edit code unless the user asked for a fix as well as investigation. Do not mutate GitHub unless explicitly authorized.

Handoff with reproducer, root cause, affected contract, proposed regression, and any sibling risks.

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.