← ClaudeAtlas

finishing-a-development-branchlisted

Use when implementation is complete and tests pass — runs the targeted test suite, then pushes the branch and opens a PR against the base branch. Creates the PR only; it never merges and never discards work.
minhtran3124/agent-harness · ★ 3 · AI & Automation · score 59
Install: claude install-skill minhtran3124/agent-harness
# Finishing a Development Branch ## Overview Complete development work by verifying tests, then opening a pull request. **Core principle:** Verify tests → Push → Open PR → Report URL. This skill **only creates a PR**. It never merges, never force-pushes, and never discards work — a human reviews and merges the PR. **Announce at start:** "I'm using the finishing-a-development-branch skill to open a PR for this work." ## The Process ### Step 1: Verify Tests Invoke the **test-runner** sub-agent to run the test suite. Pass it the list of files changed on the current branch so it can target the minimal relevant test set. #### 1a. Identify changed files ```bash git diff --name-only $(git merge-base HEAD main)...HEAD -- '*.py' ``` Use these paths to determine which test files to run (e.g. changes to `app/repositories/user.py` → run `tests/repositories/test_user.py`). #### 1b. Run tests Launch the **test-runner** sub-agent with the project's test runner. Use whatever this repo actually uses — detect it, don't assume a stack. For **this** repo (harness-skills) that is: ```bash bash scripts/run-tests.sh ``` For an application repo, run the targeted test files with the project's runner (e.g. `pytest <test_files> -x`, `npm test`, `go test ./...`); fall back to the full suite when no targeted subset applies. #### 1c. Report results Present a structured summary: ``` Test Report ─────────────────────────────── Result: ✅ N passed / ❌ N failed Files: tests/path/to/test_fi