agentleFS
Sequela02Sequela02/canasta/skills/workflow/writing-issues/SKILL.md

writing-issues

Skill0 starsChanged yesterdayMITView on GitHub

Draft or revise an issue, bug report, or feature request with

2 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. About 0 are common, 3 are distinctive, and 1 is 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.

  • 1

    Very specific

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

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

    Use for issue titles and bodies.

    line 14
  2. Distinctive:

    Read the request, relevant evidence, repository contribution rules, and the applicable issue template.

    line 18
  3. Distinctive:

    When breaking a plan into several issues, write one per independently verifiable slice (see implementation-planning), describe the …

    line 36
  4. Distinctive:

    Follow the repository template, including instructions to omit inapplicable sections.

    line 42

≈ “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

Writing Issues

Use for issue titles and bodies. Use technical-writing for prose and examples; use problem-clarification only when missing information materially affects the issue. Writing a draft does not require implementing the proposed solution.

  1. Read the request, relevant evidence, repository contribution rules, and the applicable issue template. Check related issues and work when accessible; distinguish a duplicate from a related problem. Do not invent tracker IDs, labels, priority, links, or checks. State an unavailable material source once.
  2. Name the problem or desired capability in the title. Open with who or what is affected, the trigger or unmet need, and why it matters. Describe current and desired behavior in terms someone can observe without this conversation.
  3. For a bug, give the smallest known reproduction, relevant environment or revision, expected versus actual result, and supporting evidence. Say when the report has not been reproduced. Keep suspected causes and proposed fixes separate from observations; a plausible explanation is not a diagnosis.
  4. For a feature, explain the use case and constraints. Include a proposed solution only when useful and label its decision status. Do not turn an initial request into a committed architecture or an implementation inventory.
  5. State observable acceptance criteria that distinguish success from the current problem. For an investigation, name the question and evidence needed to resolve it instead of promising a fix. Include scope boundaries, dependencies, or open decisions only when they prevent a material misunderstanding.
  6. When breaking a plan into several issues, write one per independently verifiable slice (see implementation-planning), describe the end-to-end behavior it delivers rather than a layer inventory, and list the issues that block it, using the tracker's native blocking link where it has one. Show the breakdown before creating anything; approving it does not authorize publishing or labels.
  7. Follow the repository template, including instructions to omit inapplicable sections. Without a template, a small issue can be one paragraph plus an acceptance sentence. Use numbered reproduction steps or a comparison when helpful; omit empty headings and generic checklists. Preserve necessary facts while removing credentials, private data and internal infrastructure names (secret stores, private repositories, hosts) from publication-bound evidence.
  8. Recheck that the title and body describe the same problem, links exist, and acceptance criteria are requirements rather than claims of completed work. Return the draft, or use the existing authorized tracker workflow if creation or updating was requested. Never infer publication authority from this skill.

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.