finishing-work
Use when completing work on a task or feature. Prompts for code review before finishing. ALWAYS invoke this skill before saying work is done.
When & Why to Use This Skill
The finishing-work Claude skill serves as a critical quality assurance gate for software development workflows. It ensures that no task or feature is marked as complete without a mandatory prompt for code review, integrating seamlessly with automated review pipelines to identify bugs, improve code maintainability, and enforce high development standards before final submission.
Use Cases
- Standardizing the 'Definition of Done': Enforcing a consistent workflow where every completed task must be vetted through a code review process.
- Automated Quality Pipelines: Orchestrating a multi-step review sequence (request, triage, and fix) using tools like CodeRabbit to streamline feedback loops.
- Risk Mitigation: Preventing the introduction of technical debt or regressions by ensuring even small changes are offered for peer or automated inspection.
- Developer Productivity: Reducing manual overhead by automating the transition from 'work in progress' to 'under review' with a single command.
| name | finishing-work |
|---|---|
| description | Use when completing work on a task or feature. Prompts for code review before finishing. ALWAYS invoke this skill before saying work is done. |
Finishing Work
CRITICAL: Always Ask About Code Review
Before finishing ANY work, you MUST ask the user:
"Would you like me to request a code review before finishing? (Yes/No)"
Use the AskUserQuestion tool with options:
- Yes - Invoke
coderabbit-requestskill, then triage and fix any issues - No - Skip review and proceed to finish
Do NOT skip this prompt. Even for small changes, the user decides.
If User Says Yes: Code Review Workflow
Invoke coderabbit-request to start the review pipeline:
coderabbit-request
↓ (outputs JSON)
coderabbit-triage
↓ (outputs task plan)
coderabbit-fix (parallel or sequential)