agentleFS
AgnosticUIAgnosticUI/agnosticui/.claude/skills/implement-github-feature/SKILL.md

implement-github-feature

Skill833 starsChanged 9 months agoApache-2.0View on GitHub

Design and implement a new feature or component from a GitHub issue. Use for net-new functionality, not bug fixes.

  • Commits and pushes

The file

Show its rules and text

We split this skill into the rules it gives an agent

18 rules. About 0 are common, 15 are distinctive, and 3 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈15

    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:

    If either condition fails:

    line 36
  2. Distinctive:

    Ask the user to resolve before continuing

    line 38
  3. Very specific:

    Read their implementations in v2/lib/src/components/

    line 72
  4. Distinctive:

    If scope appears ambiguous:

    line 86
  5. Distinctive:

    Ask the user before proceeding

    line 87
  6. Distinctive:

    Do not create files

    line 119
  7. Distinctive:

    Do not write code

    line 120
  8. Distinctive:

    Add placeholder files where appropriate

    line 132
  9. Distinctive:

    Add minimal README or doc stubs if needed

    line 133
  10. Distinctive:

    No implementation logic yet

    line 134
  11. Very specific:

    Run npm run build in v2/lib/

    line 158
  12. Very specific:

    Run npm run lint if available

    line 160
  13. Distinctive:

    Ensure consistency across frameworks

    line 185
  14. Distinctive:

    Update or add playbooks if applicable

    line 197
  15. Distinctive:

    Verify examples compile and behave correctly

    line 200
  16. Distinctive:

    Keep examples minimal and educational

    line 201
  17. Distinctive:

    Ensure no accidental breaking changes

    line 213
  18. Distinctive:

    Do not assume intent — if unclear, ask

    line 256

≈ “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

Usage: /implement-github-feature ISSUE_NUMBER

Example: /implement-github-feature 281

Implement GitHub feature request $ARGUMENTS following AgnosticUI conventions, accessibility standards, and CSS-first principles.


Setup

  1. Read project context
    • Read .claude/PROJECT_CONTEXT.md
    • Understand:
      • Repository structure
      • Design principles
      • Accessibility requirements
      • Branch conventions
      • Component workflows

Phase 0: Safety Checks (Required)

  1. Verify clean starting state
    • Run git status
    • Confirm:
      • Working directory is clean
      • Current branch is master
    • If either condition fails:
      • STOP
      • Ask the user to resolve before continuing

Phase 1: Branching

  1. Propose feature branch
    • Follow convention: issue-$ARGUMENTS/short-descriptive-name
    • Example: issue-281/selection-card-group
    • Run:
      git checkout -b issue-$ARGUMENTS/short-description
    • WAIT FOR USER APPROVAL of branch name before proceeding

Phase 2: Issue Analysis & Intent Extraction (No Code)

  1. Analyze the GitHub issue
    • Fetch full details:
      gh issue view $ARGUMENTS
    • Extract and summarize:
      • Feature goals
      • Non-goals / constraints
      • API expectations
      • Accessibility requirements
      • Styling and theming expectations
      • Framework implications (Lit / React / Vue)
  2. Study existing related components
    • Identify components this feature will compose or resemble
    • Read their implementations in v2/lib/src/components/
    • Note patterns for:
      • Props and attributes
      • Slot usage
      • CSS custom properties
      • ::part exposure
      • Event dispatching
      • Accessibility implementation
    • This ensures consistency with established AgnosticUI conventions
  3. Confirm feature classification
    • This is a:
      • Net-new component OR
      • Major enhancement (not a bug fix)
    • If scope appears ambiguous:
      • Ask the user before proceeding

Phase 3: Design Proposal (Hard Stop Before Code)

  1. Propose high-level design
    • Identify:
      • New components to be created
      • Public vs internal APIs
      • Required props vs optional props
      • Slot usage and responsibilities
      • Accessibility model (labels, roles, keyboard behavior)
      • Shadow DOM strategy (::part, ::slotted)
    • Explicitly call out:
      • Web Component constraints
      • CSS-first decisions
      • What is intentionally not supported
  2. Propose file & directory layout
    • Core components:
      • v2/lib/src/components/...
    • Documentation:
      • v2/site/docs/components/...
    • Storybook playgrounds:
      • Lit
      • React
      • Vue
    • Examples:
      • Lit / React / Vue test examples
  3. WAIT FOR USER APPROVAL
  • Do not create files
  • Do not write code
  • Proceed only after explicit approval of:
    • API shape
    • Accessibility approach
    • Styling strategy

Phase 4: Scaffolding (Minimal, Intentional)

  1. Create initial scaffolding
    • Create component directories
    • Add placeholder files where appropriate
    • Add minimal README or doc stubs if needed
    • No implementation logic yet
  2. Show scaffolding diff
    • Use git diff
    • Explain what was created and why
    • WAIT FOR USER APPROVAL

Phase 5: Core Implementation (Lit Web Components First)

  1. Implement core component(s)
    • Work only in:
      • v2/lib/src/components/
    • Follow:
      • CSS-first approach
      • Design token usage
      • Accessibility requirements
    • Ensure:
      • Keyboard interaction works
      • Labels and semantics are correct
      • Shadow DOM parts are intentional and minimal
  2. Verify build
    • Run npm run build in v2/lib/
    • Fix any compilation errors
    • Run npm run lint if available
  3. Pause for review
    • Explain:
      • DOM structure
      • Slot behavior
      • ::part surface
      • Accessibility decisions
    • WAIT FOR USER APPROVAL before proceeding

Phase 6: Framework Integrations

  1. Integrate with frameworks
    • React playground
    • Vue playground
    • Lit playground
    • APIs should feel idiomatic but map 1:1 conceptually
  2. Add Storybook stories
    • Demonstrate:
      • Core use cases
      • Variants
      • Accessibility states
    • Ensure consistency across frameworks

Phase 7: Documentation & Examples

  1. Write documentation
    • Add VitePress docs:
      • API
      • Slots vs props
      • Styling guidance
      • Accessibility notes
    • Update or add playbooks if applicable
  2. Update examples
    • Verify examples compile and behave correctly
    • Keep examples minimal and educational

Phase 8: Final Review & Commit

  1. Final verification
    • Review:
      • git diff
      • Accessibility behavior
      • Styling hooks
      • API consistency
    • Ensure no accidental breaking changes
  2. Prepare commit
    • Stage changes:
      git add .
    • Commit message format:
      Add #$ARGUMENTS: [concise feature description]
    • Show commit contents
    • WAIT FOR USER APPROVAL before committing

Phase 9: Handoff

  1. Explain next steps
    • User is on branch: issue-$ARGUMENTS/...
    • Review with:
      git diff master
    • Push when ready:
      git push -u origin issue-$ARGUMENTS/...
    • Create PR:
      gh pr create --base master --head issue-$ARGUMENTS/...

Important Rules

  • NEVER work directly on master
  • NEVER push without explicit user permission
  • ALWAYS stop at design gates
  • ALWAYS prioritize accessibility over convenience
  • Prefer slots for layout, props for semantics
  • CSS custom properties > ::part > ::slotted
  • Do not assume intent — if unclear, ask

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.