agentleFS
mrveissmrveiss/AutoBot-AI/.claude/skills/implement/SKILL.md

implement

Skill39 starsChanged 18 days agoApache-2.0View on GitHub

Complete GitHub issue implementation workflow with built-in guardrails

  • Commits and pushes

191 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use setup-ts-deep-modules, ask-matt 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.
  • liufeng1976Skill

    setup-ts-deep-modules

    Wire dependency-cruiser into a TypeScript repo so each package is a deep module — implementation hidden in subfolders, reachable only through its entry-point files. User-invoked.

    Also in 43 projects that use this

  • GoogleSkill

    ask-matt

    Ask which skill or flow fits your situation. A router over the skills in this repo.

    Also in 91 projects that use this

  • RoentekSkill

    loop-me

    Grill me about specs for the workflows I want to build, within this workspace.

    Also in 50 projects that use this

  • AWSSkill

    to-questionnaire

    Turn a decision you can't fully answer into a questionnaire for someone else to fill in.

    Also in 88 projects that use this

  • ryotaroh180105Skill

    claude-handoff

    Hand the current conversation off to a fresh background agent that picks up the work immediately.

    Also in 42 projects that use this

  • GoogleSkill

    to-spec

    Turn the current conversation into a spec and publish it to the project issue tracker — no interview, just synthesis of what you've already discussed.

    Also in 168 projects that use this

  • +3 more from google

The file

Show its rules and text

We split this skill into the rules it gives an agent

27 rules. 0 are common, 24 are distinctive, and 3 are very specific to this project.

  • 0

    Common

    Most of the 171 edited versions we read keep these.

  • 24

    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:

    Before making ANY code changes:

    line 12
  2. Distinctive:

    If on wrong branch, STOP and ask user

    line 20
  3. Distinctive:

    If uncommitted changes exist, STOP and ask user how to handle

    line 21
  4. Distinctive:

    If stashes exist, STOP and ask user how to handle

    line 22
  5. Distinctive:

    Verify issue is still open

    line 28
  6. Distinctive:

    Read full description and acceptance criteria

    line 29
  7. Distinctive:

    Check for linked PRs or existing branches

    line 30
  8. Distinctive:

    If PR/branch exists, STOP and ask user

    line 37
  9. Very specific:

    Use mcp__memory__search_nodes to find relevant past work

    line 54
  10. Distinctive:

    Check for related issues or patterns

    line 55
  11. Distinctive:

    Add subtasks as checklist to GitHub issue

    line 59
  12. Very specific:

    Use Edit for existing files >50 lines (NOT Write)

    line 64
  13. Distinctive:

    Run relevant tests after each subtask

    line 66
  14. Distinctive:

    Commit frontend when complete

    line 80
  15. Distinctive:

    Run full test suite

    line 81
  16. Distinctive:

    Run git diff --staged before committing

    line 95
  17. Distinctive:

    Verify no unintended files are staged

    line 96
  18. Distinctive:

    If pre-commit hooks fail, fix the issue (don't bypass)

    line 97
  19. Very specific:

    Use mcp__memory__create_entities to record:

    line 156
  20. Distinctive:

    Commit completed work immediately

    line 165
  21. Distinctive:

    Ask user if should continue or pause

    line 167
  22. Distinctive:

    Document blocker in issue comment

    line 170
  23. Distinctive:

    Ask user how to proceed

    line 171
  24. Distinctive:

    Don't implement workarounds without permission

    line 172
  25. Distinctive:

    Create new issue for additional work

    line 175
  26. Distinctive:

    Ask user if should include in current PR or defer

    line 176
  27. Distinctive:

    Don't skip tests or mark as "will fix later"

    line 180

“Common” is measured: kept by at least 6 in 10 of the 171 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

Implement GitHub Issue

Complete workflow for implementing a GitHub issue from investigation through PR creation, merge, and cleanup.

Pre-Flight Checks (MANDATORY - run first)

Before making ANY code changes:

  1. Verify branch state:
    git branch --show-current
    git status
    git stash list
    • If on wrong branch, STOP and ask user
    • If uncommitted changes exist, STOP and ask user how to handle
    • If stashes exist, STOP and ask user how to handle
  2. Fetch issue details:
    gh issue view {issue_number}
    • Verify issue is still open
    • Read full description and acceptance criteria
    • Check for linked PRs or existing branches
  3. Check for existing work:
    gh pr list | grep {issue_number}
    git branch -a | grep {issue_number}
    • If PR/branch exists, STOP and ask user
  4. Verify target merge branch:
    git fetch origin main
    git log --oneline origin/main -3
    • Confirm main is the target (unless explicitly told otherwise)

Implementation Phase

  1. Create feature branch:
    git checkout -b fix/issue-{issue_number} main
  2. Search memory for context:
    • Use mcp__memory__search_nodes to find relevant past work
    • Check for related issues or patterns
  3. Create implementation plan:
    • Break work into 3-10 subtasks
    • Add subtasks as checklist to GitHub issue
    • Identify backend vs frontend work
    • Plan commit checkpoints (every 2-3 logical tasks)
  4. Implement using incremental edits:

    • Use Edit for existing files >50 lines (NOT Write)
    • Focus on one subtask at a time
    • Run relevant tests after each subtask
    • Commit after completing each logical group (2-3 subtasks)

    Commit pattern:

    git add <files-for-this-group>
    git commit -m "<type>(scope): <description> (#{issue_number})"
    git log -1 --stat  # Verify commit landed
    git diff           # Verify nothing left uncommitted
  5. For multi-phase work (backend + frontend):
    • Complete and commit backend FIRST
    • Then start frontend
    • Commit frontend when complete
    • Run full test suite

Verification Phase

  1. Run tests:
    # Backend
    python -m pytest {affected_test_files}
    
    # Frontend (if applicable)
    npm run test {affected_components}
  2. Pre-commit validation:
    • Run git diff --staged before committing
    • Verify no unintended files are staged
    • If pre-commit hooks fail, fix the issue (don't bypass)

PR & Merge Phase

  1. Push and create PR:
    git push -u origin fix/issue-{issue_number}
    
    gh pr create \
      --base main \
      --title "fix: <clear title> (#{issue_number})" \
      --body "$(cat <<'EOF'
    ## Summary
    Fixes #{issue_number}
    
    <1-3 bullet points summarizing changes>
    
    ## Changes
    - <specific change 1>
    - <specific change 2>
    
    ## Testing
    - [ ] Backend tests pass
    - [ ] Frontend tests pass (if applicable)
    - [ ] Manual testing completed
    
    🤖 Generated with Claude Code
    EOF
    )"
  2. Merge PR:
    # Verify PR is approved and CI passes
    gh pr merge {pr_number} --squash --delete-branch
  3. Cleanup:
    git checkout main
    git pull origin main
    git branch -d fix/issue-{issue_number}
  4. Close issue — run the closure gates first (CLAUDE_CLOSURE.md):
    # Gate 2: forward-tracking references block closure (exit 1 → do NOT close)
    pipeline-scripts/check-issue-close-refs.sh {issue_number}
    
    gh issue close {issue_number} --comment "Implemented in PR #{pr_number}. Gate 2 clear.
    
    Acceptance criteria (quoted verbatim from issue body):
    - \"<exact AC text>\" — ✅ <evidence>
    - \"<exact AC text>\" — ❌ NOT met → follow-up #XXXX"
    • Gate 1: quote each issue-body AC verbatim — never paraphrase what shipped
    • Gate 3: if this PR bundled several issues, close {issue_number} only if its FULL AC set is delivered; a subset → check off the subtask and keep it open (or name the follow-up in the same comment)
  5. Store in memory:
    • Use mcp__memory__create_entities to record:
      • What was implemented
      • Key technical decisions
      • Patterns used
      • Lessons learned

Guardrails

If session is getting long:

  • Commit completed work immediately
  • Note what's left in issue comment
  • Ask user if should continue or pause

If implementation is blocked:

  • Document blocker in issue comment
  • Ask user how to proceed
  • Don't implement workarounds without permission

If scope expands:

  • Create new issue for additional work
  • Ask user if should include in current PR or defer

If tests fail:

  • Fix failures before creating PR
  • Don't skip tests or mark as "will fix later"

Success Criteria

✅ All code committed with correct issue references ✅ All acceptance criteria verified ✅ Tests passing ✅ PR created and merged to main ✅ Feature branch deleted ✅ Issue closed with summary ✅ Knowledge stored in memory MCP

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.