agentleFS
holon-runholon-run/holon/skills/github-issue-solve/SKILL.md

github-issue-solve

Skill194 starsChanged 4 days agoApache-2.0View on GitHub

Solve a GitHub issue by collecting context, implementing a fix, and opening or updating a pull request.

  • Reads credentials
  • Commits and pushes

The file

Show its rules and text

We split this skill into the rules it gives an agent

14 rules. About 0 are common, 10 are distinctive, and 4 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈10

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 4

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Use this skill when you need to turn a GitHub issue into a concrete code change and publish the result as a pull request.

    line 10
  2. Distinctive:

    Do not create fixed output files unless the caller or an integration explicitly requires them.

    line 33
  3. Very specific:

    If no manifest is provided, collect issue metadata and comments directly with gh:

    line 44
  4. Very specific:

    Do not assume fixed file names under github/.

    line 51
  5. Very specific:

    If ${GITHUB_CONTEXT_DIR}/manifest.json exists, use it.

    line 58
  6. Distinctive:

    Follow the developer agent's recorded worktree, branch, and PR preferences.

    line 65
  7. Distinctive:

    Run relevant verification commands before publish.

    line 67
  8. Distinctive:

    Commit only intentional changes.

    line 71
  9. Very specific:

    Use raw gh commands. Pass the PR body directly, or use a caller-requested body file when the integration needs one:

    line 76
  10. Distinctive:

    Keep scope aligned with issue intent.

    line 94
  11. Distinctive:

    If full verification is impossible, report what was attempted and why it is incomplete.

    line 97
  12. Distinctive:

    no meaningful code change was produced for the issue intent

    line 102
  13. Distinctive:

    commit/push was not completed

    line 103
  14. Distinctive:

    Do not report success from artifacts alone.

    line 106

≈ “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

GitHub Issue Solve Skill

Summary

Use this skill when you need to turn a GitHub issue into a concrete code change and publish the result as a pull request.

When To Use

  • Fixing a GitHub issue end-to-end
  • Collecting issue context and comments with raw gh commands
  • Implementing code changes and opening or updating a PR

Do Not Use

  • Reviewing an existing PR without making changes
  • Replying to review feedback on an existing PR
  • Project-wide planning or backlog triage

Prerequisites

  • gh CLI authentication is required.
  • GITHUB_TOKEN/GH_TOKEN must allow issue/PR read-write operations.

Runtime Paths

  • GITHUB_OUTPUT_DIR: optional caller-provided output artifacts directory.
  • GITHUB_CONTEXT_DIR: context directory (default ${GITHUB_OUTPUT_DIR}/github-context).
  • Do not create fixed output files unless the caller or an integration explicitly requires them.

Inputs (Manifest-First)

Preferred input when already available:

  • ${GITHUB_CONTEXT_DIR}/manifest.json

Optional inputs:

  • Any artifact listed as status=present in manifest.artifacts[].

If no manifest is provided, collect issue metadata and comments directly with gh:

gh issue view <issue_number> --repo <owner/repo> --json number,title,body,state,url,author,createdAt,updatedAt,labels
gh api repos/<owner>/<repo>/issues/<issue_number>/comments --paginate

Do not assume fixed file names under github/. Resolve usable inputs from manifest.artifacts[] by id/path/status/description.

Workflow

1. Collect context

  • If ${GITHUB_CONTEXT_DIR}/manifest.json exists, use it.
  • Otherwise, collect the issue body and comments directly with gh.

2. Analyze and implement

  • Extract acceptance criteria and constraints from issue metadata and discussion.
  • Implement minimal complete changes for the requested outcome.
  • Follow the developer agent's recorded worktree, branch, and PR preferences; ask the operator when a relevant preference is not recorded.
  • Run relevant verification commands before publish.

3. Commit and push

  • Commit only intentional changes.
  • Push branch to remote before PR publish.

4. Publish PR

Use raw gh commands. Pass the PR body directly, or use a caller-requested body file when the integration needs one:

gh pr create --repo <owner/repo> --title "<title>" --body "<body>" --head <branch> --base <base>
gh pr edit <pr_number> --repo <owner/repo> --title "<title>" --body "<body>"

Publish completion is mandatory; do not report success without a real PR side effect.

5. Finalize delivery

Report the result in the normal agent delivery. If a caller explicitly requires machine-readable output, write only the requested artifacts under ${GITHUB_OUTPUT_DIR} and include their paths in the delivery.

Delivery Standards

  • Keep scope aligned with issue intent; avoid unrelated refactors.
  • State assumptions explicitly when requirements are ambiguous.
  • Include concrete verification results (commands + outcomes).
  • If full verification is impossible, report what was attempted and why it is incomplete.

Failure Rules

Mark run as failed if any of the following is true:

  • no meaningful code change was produced for the issue intent
  • commit/push was not completed
  • PR create/update failed or PR URL cannot be verified

Do not report success from artifacts alone. A summary or manifest is optional evidence, not a substitute for the actual commit, push, or PR side effect.

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.