Pull Request
September 11, 2026 Β· View on GitHub
π First time contributing? Read both:
- CONTRIBUTING.md Β§ Before You Submit β the contract and the workflow
- Plugin & Skill Spec (v3.6.0 baseline; enforced schema 3.15.2 β see
000-docs/SCHEMA_CHANGELOG.md) β the authoritative spec the validator scores against (8-field frontmatter, 7 body sections, 100-point rubric, all source-cited)The PR Pre-screen workflow sets an advisory
prescreen-gradestatus on every run and posts a structured comment with the exact validator findings only when changes are needed (silent on PASS) β reading both up front saves a round-trip.
Type of Change
- π New plugin submission
- β¬οΈ Plugin update/enhancement
- π Documentation improvement
- π Bug fix
- ποΈ Infrastructure/CI improvement
- π¨ Marketplace website update
- π§ Configuration change
- π§ͺ Tests added/updated
- π Security fix
- π Performance improvement
- β»οΈ Refactoring
- π Translation/i18n
- ποΈ Deprecation/removal
- π¦ Dependency update
- π― Other (please describe)
Description
Summary:
Motivation:
Related Issues:
Submission Standard (plugin & source submissions)
- This PR links its submission issue (
Closes #N/Refs #N) β submissions start as an issue using the plugin-submission template; a PR without one will be asked to file it - The required docs for my declared tier are included (
docs/PRD.mdat minimum β see the tierβdocs matrix) - Validator run locally (
python3 scripts/validate-skills-schema.py --marketplace --verbose <plugin-dir>) - For
feat(sources)PRs: source vetted per the external-source vetting playbook -
verified:/curated:flags insources.yamlare honest (noverified: truewithout maintainer vetting; no silently hardened mirrors withoutcurated: true)
Security
- Related alerts: GHSA-/CVE-
- Impact assessment: None / Low / Medium / High
Plugin Details (for plugin submissions/updates)
Plugin Name: Category: Version: Keywords:
Components Included:
- Commands (how many: )
- Agents (how many: )
- Hooks (how many: )
- Scripts (how many: )
- MCP servers
Dependencies:
Checklist
For All PRs
- I have read the CONTRIBUTING.md guidelines
- "Allow edits by maintainers" is enabled (fork PRs) β we push small regen/bump fixes to your branch instead of round-tripping
- My code follows the project's style and conventions
- I have performed a self-review of my code
- I have commented my code where necessary
- My changes generate no new warnings or errors
- Documentation has been updated (if applicable)
- CodeQL passes; no new code scanning alerts
- No high/critical Dependabot alerts introduced
For Plugin Submissions/Updates
- Plugin has valid
.claude-plugin/plugin.jsonwith all required fields -
plugin.jsonvalidated withjq empty plugin.json - README.md is comprehensive with installation, usage, and examples
- LICENSE file is included (MIT or Apache-2.0 recommended)
- All shell scripts are executable (
chmod +x scripts/*.sh) - No hardcoded secrets, API keys, or credentials
- Marketplace.json has been updated with plugin entry
- Plugin tested locally with
/plugin install - Commands include YAML frontmatter (name, description)
- Agents include YAML frontmatter (name, description, model)
- Hooks use
${CLAUDE_PLUGIN_ROOT}for portable paths - All JSON files are valid (
jq empty *.json)
For Marketplace Website Changes
- Website builds successfully (
cd marketplace && npm run build) - No broken links or missing assets
- Mobile-responsive design verified
- SEO metadata updated (if applicable)
- Accessibility standards followed (WCAG 2.1)
For Documentation Changes
- Spelling and grammar checked
- Code examples tested and working
- Links verified and not broken
- Screenshots/diagrams updated (if applicable)
- Version references updated
Testing Evidence
Test Environment:
- OS:
- Claude Code Version:
- Python Version (if applicable):
- Node Version (if applicable):
Test Commands Run:
# Example:
/plugin install my-plugin@test
/my-command --option value
Test Results:
Edge Cases Tested:
- Works with minimal configuration
- Handles errors gracefully
- Works across different platforms (macOS, Linux, Windows)
- Performance is acceptable for large inputs
- Security considerations addressed
Breaking Changes
- Yes (describe below)
- No
If yes, describe:
Screenshots / Recordings (if applicable)
Performance Impact
- No performance impact
- Minor performance impact (negligible)
- Moderate performance impact (describe below)
- Significant performance impact (describe and justify below)
Security Considerations
Automated Security Scans
- No hardcoded secrets (API keys, passwords, tokens)
- No AWS keys, private keys, or credentials detected
- No destructive commands (
rm -rf /, data deletion) - No
eval()or command injection risks - No suspicious curl/wget to IP addresses
- No base64 obfuscation detected
- All URLs use HTTPS (except localhost)
- No URL shorteners (potential phishing)
Manual Security Review
- Prompt Injection Protection: No hidden instructions that hijack Claude's behavior
- Data Privacy: No data exfiltration to external servers (all network calls documented in README)
- Permission Audit: Plugin requests minimal necessary permissions only
- Clear Intent: README clearly explains what the plugin does
- Input Validation: All user inputs validated before processing
- Error Handling: No sensitive data in error messages
- Dependencies: All npm dependencies reviewed (for MCP plugins)
Testing
- Tested in isolated environment before submission
- No unexpected side effects observed
- Graceful error handling for edge cases
- Works as documented in README
Rollback Plan
Additional Notes
Reviewer Checklist (for maintainers)
- Code quality meets standards
- Tests pass (CI/CD green)
- Security review completed
- Documentation is clear and complete
- Breaking changes documented
- Version bumped appropriately
- CHANGELOG.md updated
- Ready to merge
By submitting this PR, I confirm:
- I have the right to submit this code under the project's license
- I understand my contributions will be publicly available
- I agree to the project's Code of Conduct