agentleFS
anorthanorth/agent-skills/skills/verify-ticket/SKILL.md

verify-ticket

Skill1 starsChanged 5 days agoMITView on GitHub

Verifies a change against a ticket's done checks as a non-author: runs each check on the best surface available, the way a user would, and returns PASS, PASS+NOTES, FAIL, or INCONCLUSIVE with evidence. Use when the user or agent asks to verify a ticket, a branch, or a built change.

1 other project uses a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

10 rules. About 0 are common, 7 are distinctive, and 3 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈7

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 3

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Only a verdict from an agent that did not write the code counts.

    line 14
  2. Distinctive:

    Read the ticket in full, with its notes and comments.

    line 23
  3. Distinctive:

    If the user names a ref instead, check it out into a temporary git worktree, verify that tree, and remove it after.

    line 25
  4. Very specific:

    If it's unclear which ticket or change, ask.

    line 26
  5. Very specific:

    If this conversation wrote or edited the change, or you can't tell, spawn a verifier subagent.

    line 35
  6. Very specific:

    Start it with a fresh context window, not a fork, on the standard tier (see /model-tiers).

    line 36
  7. Distinctive:

    Never pass the builder's done-check evidence, results, or report.

    line 47
  8. Distinctive:

    If the verifier errors or returns no verdict, empty the evidence directory and retry once.

    line 49
  9. Distinctive:

    If that fails too, report the blocker and stop.

    line 50
  10. Distinctive:

    If your model is the implementer's, say so in the verdict.

    line 53

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 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

Verify ticket

Only a verdict from an agent that did not write the code counts. A verifier works from the ticket, runs its done checks, and returns a verdict with evidence. Green tests and an approving review are not verdicts.

You coordinate: load, choose who verifies, brief, relay, record. The verifier follows verifier.md, in this skill's directory.

1. Load

Read the ticket in full, with its notes and comments. The change to verify is the working tree. If the user names a ref instead, check it out into a temporary git worktree, verify that tree, and remove it after. If it's unclear which ticket or change, ask. Find the implementer's model: an Implementer: line in the ticket's comments, else this conversation's model if it wrote the change.

The evidence directory is ${TMPDIR:-/tmp}/verify-ticket/<repo directory name>/<ticket-id>/. Empty it if it exists.

2. Choose who verifies

If this conversation wrote or edited the change, or you can't tell, spawn a verifier subagent. Start it with a fresh context window, not a fork, on the standard tier (see /model-tiers). Brief it with:

Read and follow <absolute path to this skill's directory>/verifier.md.
Ticket: <ticket id, or its full text with its notes and comments>
Change: working tree at <absolute path>
Implementer: <model, or "unknown">
Evidence directory: <absolute path>

Never pass the builder's done-check evidence, results, or report. A verdict of any kind, FAIL included, is the result. If the verifier errors or returns no verdict, empty the evidence directory and retry once. If that fails too, report the blocker and stop.

Otherwise, verify inline: follow verifier.md yourself. If your model is the implementer's, say so in the verdict.

3. Relay and record

Report the verdict verbatim. Don't soften, reinterpret, or override it. Comment on the ticket (see /ticket-states) in one line: the verdict, the verifier's model, and the evidence directory.

Change the ticket's state only when the calling skill says to.

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.