Finishing a Development Branch
Independent PiSkill directory guide. The original skill remains hosted by Superpowers.
What is Finishing a Development Branch?
Verifies completed development work and guides the final merge, pull-request, branch-cleanup, or integration decision only after tests have passed.
What does Finishing a Development Branch do?
Finishing a Development Branch is a Superpowers skill for deciding what to do with completed work after implementation and verification. It requires a passing test state first, then guides the user through merge, pull request, keeping the branch or discarding the work.
Who is Finishing a Development Branch best for?
- Developers finishing AI-assisted feature branches
- Teams standardizing branch-completion decisions
- Projects that require verification before pull request or merge
- Worktree users cleaning up completed feature environments
Common use cases
- Verify tests before declaring a branch finished
- Choose between local merge and pull request
- Keep a branch for later work without premature cleanup
- Remove worktree state after an approved completion path
How does Finishing a Development Branch work?
The workflow runs the relevant verification first and refuses to proceed when tests fail. Once the branch is healthy, it presents explicit completion options and performs the selected Git or worktree cleanup steps.
Key benefits
- Prevents branch-finalization before verification
- Makes completion options explicit
- Supports both pull-request and local-merge workflows
- Integrates cleanup with worktree-based development
Things to know
- It depends on the project's tests being meaningful
- Repository-specific branch protection and CI rules still apply
- The user still needs to decide which completion path is appropriate
Compatible tools
Frequently asked questions
What should happen before merging an AI-generated branch?
Does the skill automatically merge the branch?
Related skills
Verification Before Completion
Requires fresh evidence before an agent claims work is fixed, passing, or complete, making it especially useful for validating AI-generated code before commits or pull requests.
Using Git Worktrees
Creates or verifies an isolated Git worktree before feature implementation so agent work stays separated from the main workspace and can be developed safely in parallel.
Receiving Code Review
Helps agents evaluate code-review feedback technically before implementing it, verify suggestions against the codebase, and avoid blindly accepting incorrect or incomplete review comments.
Requesting Code Review
Structures code-review requests around the actual implementation, requirements, and git diff so issues can be caught before they reach the main branch.
Application Insights for Web
Adds Azure Application Insights to TypeScript web applications with telemetry initialization, page tracking, custom events, correlation, privacy controls, and verification.
AWS Context Discovery
Discovers the active AWS profile, region, account, and caller identity before Hugging Face cloud or SageMaker work so later steps use the correct AWS context.