From claude-skills
Creates a pull request from a feature branch into main after verifying all quality gates — closed issues, PRD acceptance criteria, test coverage, and code quality. Triggers on "create the feature PR", "open a PR for the feature", "ship this feature", or any request to finalize a feature.
How this skill is triggered — by the user, by Claude, or both
Slash command
/claude-skills:feature-pr [feature-name][feature-name]This skill is limited to the following tools:
The summary Claude sees in its skill listing — used to decide when to auto-load this skill
Create a pull request from the feature branch into main. Run all quality gates first — all issues closed, no outstanding PRs against the feature branch, PRD fully implemented, integration tests covering the full user journey, code clean. Open the PR regardless of gate results, documenting any blockers clearly in the body. Merging to main is a human decision enforced by GitHub branch protection.
Create a pull request from the feature branch into main. Run all quality gates first — all issues closed, no outstanding PRs against the feature branch, PRD fully implemented, integration tests covering the full user journey, code clean. Open the PR regardless of gate results, documenting any blockers clearly in the body. Merging to main is a human decision enforced by GitHub branch protection.
If a feature name was provided as an argument, use it. Otherwise infer from the current branch name. If neither is available, list feature branches and ask:
git branch -r | grep -v HEAD | sed 's/ origin\///' | grep -v main
Locate the PRD at docs/PRD/<feature-name>.md. If it doesn't exist, stop — this workflow requires a PRD.
Run all gates and collect results. Do not stop at the first failure — evaluate everything so the PR body captures the full picture.
gh issue list --label "<feature-name>" --state open --json number,title
Pass: no open issues. Fail: list which are open.
gh pr list --base feature/<feature-name> --state open --json number,title,headRefName
Pass: no open PRs. Fail: list which are open.
Read the PRD in full. Examine the codebase against every acceptance criterion — not as a paperwork exercise, but as a genuine code review.
For each criterion: find the code that implements it, verify the implementation satisfies the intent, note anything absent or partially implemented.
Then verify each Quality Attribute from the PRD. For each attribute, find evidence across the feature codebase that the constraint is addressed. A quality attribute that no slice addresses is a gap — note exactly which attribute and what's missing.
Evaluate whether module boundaries across all slices form a coherent design. Check for: circular dependencies between modules introduced by different slices, public APIs that leak implementation types across module boundaries, and abstractions introduced by one slice that are never reused by others.
Pass: every acceptance criterion is verifiably implemented, every quality attribute is addressed, and module boundaries are coherent. Fail: note exactly which criteria, attributes, or boundary issues and what's missing.
git checkout feature/<feature-name>
git pull
Integration coverage: Evaluate whether the test suite covers the full user journey — not just the sum of individual issue tests. Look for end-to-end coverage, error cases spanning multiple slices, and gaps that only appear when the feature is considered as a whole.
Run the full test suite. All tests must pass.
Code quality: Run the formatter and linter — must be clean.
Pass: tests pass, coverage complete, code clean. Fail: note exactly what failed.
Open the PR regardless of gate results. Use templates/feature-pr-body.md for the body structure.
gh pr create \
--title "<Feature Name>" \
--body "<filled PR body>" \
--base main \
--label "<feature-name>"
Report the PR URL and summarize any open blockers so the user knows what needs attention before merge.
npx claudepluginhub teamvoth/claude-skills --plugin claude-skillsGuides creation, editing, and verification of skills for AI coding agents using test-driven development with subagent scenarios. Use when authoring or debugging skills.