finish-issue

Runs tests, code review, PR creation, and CI verification for completed Linear issues.

Updated Nov 30, 2025
One-click install
npx skills add https://github.com/Amakaflow/amakaflow-dev-workspace --skill finish-issue-amakaflow
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: finish-issue
Source: https://github.com/Amakaflow/amakaflow-dev-workspace/tree/main/.claude/skills/finish-issue
Command: npx skills add https://github.com/Amakaflow/amakaflow-dev-workspace --skill finish-issue-amakaflow

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? It prevents developers from declaring a Linear issue done prematurely by enforcing a quality gate: tests must pass, coverage gaps and code issues must be reviewed, a PR must be created, and CI must be green before completion. ## Core Features & Use Cases - Repo-aware test execution: Runs the correct test and lint commands per repository (pytest, xcodebuild, gradlew, supabase lint, maestro). - Agent-driven review chain: Dispatches testing-strategist, code-reviewer, and optionally e2e-qa-automation subagents to catch coverage gaps, security issues, and E2E needs. - PR and CI enforcement: Creates a structured GitHub PR with test plan and Closes AMA-NNN, then watches CI runs and distinguishes real failures from transient infrastructure crashes. - Use Case: After finishing work on Linear issue AMA-653 in chat-api, invoke the skill to run pytest, get a review, open the PR, and confirm all GitHub checks pass before closing the ticket. ## Quick Start Invoke the skill with the Linear issue ID, for example /finish-issue AMA-653, while on the feature branch for that issue.

Frequently Asked Questions about finish-issue

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I finish a Linear issue with tests and PR creation?

Invoke /finish-issue with the Linear issue ID, such as /finish-issue AMA-653, while on the feature branch. The workflow runs tests, dispatches review agents, creates the PR with gh, and watches CI until all checks pass.

What test commands run for each repository?

chat-api uses pytest, the iOS app uses xcodebuild test, the Android app uses ./gradlew test, amakaflow-db uses supabase db lint, and amakaflow-automation uses maestro test. Other repos default to pytest or their CLAUDE.md instructions.

Can I create the PR before tests pass?

No. The workflow explicitly blocks PR creation until local tests and linting pass. Failing tests must be fixed first, and the testing-strategist and code-reviewer agents run before the PR is opened.

What happens when GitHub CI fails after the PR is created?

The workflow reads the failure logs with gh run view --log-failed, fixes the root cause locally, re-runs tests, pushes the fix, and watches the new run. Transient infrastructure errors like SDK crashes or timeouts are handled with gh run rerun --failed.

When is the e2e-qa-automation agent invoked?

It runs only when the issue touches user-facing flows, such as new endpoints, changed request or response shapes, auth changes, or new UI screens. Purely internal changes skip this optional step.