agentleFS
nshopiknshopik/dream-team/skills/triage-issue/SKILL.md

triage-issue

Skill2 starsChanged 19 days agoMITView on GitHub

Triages one tracker issue before it is built: checks whether the code already does what it asks, reproduces a bug claim, asks the reporter for facts only they know and parks the issue on `workflow::blocked`, puts design and scope choices to the user, and rewrites the body in scope-issue form with the original report kept at the end. On an issue with `workflow::needs-brainstorm`, runs brainstorm's question rounds and rewrites its proposal. With no argument, lists the untriaged open issues, the blocked ones whose reporter replied and the needs-brainstorm ones. Use when the user asks to triage, check or vet an issue, a legacy issue or one filed by someone outside the project — e.g. "triage #42", "triage-issue", "what needs triage" — or when build-issue stops and names triage-issue. Not for building the issue (use build-issue) or filing a new one (use scope-issue).

48 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use prd-to-plan, prd-to-issues and 10 more.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • HK-hubSkill

    prd-to-plan

    Turn a PRD into a multi-phase implementation plan using tracer-bullet vertical slices, saved as a local Markdown file in ./plans/. Use when user wants to break down a PRD, create an implementation plan, plan phases from a PRD, or mentions "tracer bullets".

    Also in 11 projects that use this

  • kennethkeimSkill

    prd-to-issues

    Break a PRD into independently-grabbable GitHub issues using tracer-bullet vertical slices. Use when user wants to convert a PRD to issues, create implementation tickets, or break down a PRD into work items.

    Also in 13 projects that use this

  • simo97Skill

    request-refactor-plan

    Use this skill when user wants to create a refactor plan.

    Also in 12 projects that use this

  • kennethkeimSkill

    write-a-prd

    Create a PRD through user interview, codebase exploration, and module design, then submit as a GitHub issue. Use when user wants to write a PRD, create a product requirements document, or plan a new feature.

    Also in 13 projects that use this

  • simo97Skill

    design-an-interface

    Generate multiple radically different interface designs for a module using parallel sub-agents. Use when user wants to design an API, explore interface options, compare module shapes, or mentions "design it twice".

    Also in 11 projects that use this

  • DockerSkill

    fix-issue

    Fix a single GitHub issue end-to-end: triage, research, write the fix, review, and create a PR. Use when asked to fix an issue: "fix issue 1234", "resolve #500", "create a PR for issue 200".

    Also in 6 projects that use this

  • +1 more from kennethkeim
  • +3 more from simo97

The file

We split this skill into the rules it gives an agent

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

  • 0

    Common

    Most of the 47 edited versions we read keep these.

  • 8

    Distinctive

    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:

    Never start the build.

    line 21
  2. Distinctive:

    Create no label, and write no AI disclaimer line.

    line 22
  3. Very specific:

    Add the open issues with workflow::blocked whose author commented after the last triage notes.

    line 53
  4. Very specific:

    Add the open issues with workflow::needs-brainstorm.

    line 55
  5. Distinctive:

    Ask which one to triage, then run the steps below on it.

    line 56
  6. Distinctive:

    Read every earlier triage notes comment and the replies after it.

    line 66
  7. Distinctive:

    Never re-ask a question a reply answered.

    line 67
  8. Distinctive:

    Search the code for the requested behaviour by what it does, not the issue's wording.

    line 72
  9. Distinctive:

    Read the commits since the issue opened for a change that did it.

    line 73
  10. Very specific:

    No open item is left in its proposal → remove workflow::needs-brainstorm.

    line 116
  11. Distinctive:

    Leave out a line with nothing to report.

    line 128

“Common” is measured: kept by at least 6 in 10 of the 47 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

triage-issue

One issue per invocation.

  • Never start the build.
  • Create no label, and write no AI disclaimer line.

Untriaged

An open issue is untriaged when its body has no collapsed Original report block and either holds:

  • Its body is in none of: the shape scope-issue gives its kind; the older problem, proposal and Done when … paragraph; the older context, problem, proposal.
  • Its author is outside the project.

Outside the project, by forge permission:

  • GitHub: an author_association other than OWNER, MEMBER or COLLABORATOR (gh api repos/<owner>/<repo>/issues/<n> --jq .author_association).
  • GitLab: an access level below Developer (30) (glab api projects/<p>/members/all/<author id> | jq .access_level; a 404 is none).

An outside author's body counts its Original report block only when a project member made the last body edit:

  • GitHub: the issue's editor is set and is not the author; only collaborators edit another user's issue: gh api graphql -f query='{repository(owner:"<o>",name:"<r>"){issue(number:<n>){editor{login}}}}'
  • GitLab: the issue's lastEditedBy has Developer access or above, by the check above: glab api graphql -f query='{project(fullPath:"<p>"){issue(iid:"<n>"){lastEditedBy{id}}}}'

Queue: no argument

  • List the open untriaged issues, oldest first, leaving out those with workflow::blocked and triage notes.
  • Add the open issues with workflow::blocked whose author commented after the last triage notes.
  • Add the open issues with workflow::needs-brainstorm.
  • Ask which one to triage, then run the steps below on it.
- **#<n>** <title> — <not in scope-issue form | outside author | reporter replied | needs brainstorm>

1. Read

  • Fetch title, body, labels and comments per <base directory>/../build-issue/SKILL.md step 1.
  • Triaged, with no triage notes and no workflow::needs-brainstorm → report it ready and stop.
  • Read every earlier triage notes comment and the replies after it.
  • Never re-ask a question a reply answered.
  • The reporter answered every question → remove workflow::blocked.

2. Already done

  • Search the code for the requested behaviour by what it does, not the issue's wording.
  • Read the commits since the issue opened for a change that did it.
  • Found → close the issue with a comment naming the file, symbol or commit where it lives; stop.

3. Bug claim

  • The issue reports a bug → reproduce it from the reporter's steps on the default branch, locally first.
  • It needs a lab host the project's CLAUDE.md or AGENTS.md names → run it through dream-team:lab-runner.
  • Record one verdict: confirmed, with the code path; not reproduced; or not enough detail.

4. Facts only the reporter knows

  • A missing step, version or environment, or a bug not reproduced → post the triage notes comment, add workflow::blocked, leave the body unchanged, and stop.
  • Each question names the exact command output, version, setting or step wanted.
## What we've established so far
- <fact, and how it was checked>

## What we still need from you
- <question>

5. Design and scope choices

  • Load scope-issue.
  • No workflow::needs-brainstorm, a design or scope choice open → settle it with the user per its Open questions, as for an issue the user asked for in chat, before the rewrite.
  • The issue has workflow::needs-brainstorm → run dream-team:brainstorm on its proposal's open items, skipping its path proposal and hand-off; the issue is the path.
  • The user rejects the request → close the issue as not planned with a comment giving the reason, add wontfix, write no repo file, and stop.

6. Rewrite

  • Rewrite the body in scope-issue form, carrying the verified facts, the bug verdict and the user's answers.
  • End the body with the reporter's text, verbatim, in a collapsed block.
  • The body already has that block → keep it as it is.
  • The issue has workflow::needs-brainstorm → rewrite its proposal per scope-issue Updates with the user's answers, in place of the three bullets above.
  • No open item is left in its proposal → remove workflow::needs-brainstorm.
<details><summary>Original report</summary>

<reporter's text, verbatim>

</details>

Report

  • Leave out a line with nothing to report.
- **Issue:** #<n> <title>
- **Outcome:** closed, already done | closed, rejected | waiting on the reporter | needs brainstorm | ready
- **Searched:** <paths, symbols and commit ranges the already-done check read>
- **Bug:** confirmed at <code path> | not reproduced | not enough detail
- **Still open:** <open items left in the proposal; `workflow::needs-brainstorm` stays>
- **Next:** `build-issue #<n>`

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.