From dev-pipeline
Wrap up after implementation: cleanup, dev_checks, e2e, changelog, PR ready, monitor feedback, wait for merge. Use when joining mid-flow after implementation. /dev does this automatically.
How this skill is triggered — by the user, by Claude, or both
Slash command
/dev-pipeline:finish-ticketThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Takes over from "implementation complete" and drives through to merge. The human makes the merge decision. Once merged, closes out tracking. Use this when joining a session mid-flow. `/dev` runs this automatically as part of the full pipeline.
Takes over from "implementation complete" and drives through to merge. The human makes the merge decision. Once merged, closes out tracking. Use this when joining a session mid-flow. /dev runs this automatically as part of the full pipeline.
Commit early and often. Push after every commit.
style: fix lint/formatting)Run pwd. Must contain .claude/worktrees/.
If not: STOP with error.
dev/OBJECTIVE.md for the goal, tracking info, and Trello referenceRun unit tests: uv run pytest tests/unit_tests/ -v
If tests fail: fix them. Do not proceed until unit tests pass. Commit and push any fixes.
Invoke the /cleanup skill to fix code quality issues. Wait for completion.
Commit and push cleanup fixes immediately (style: cleanup pass).
Execute ./scripts/dev_checks.sh.
./scripts/format_all.sh, commit and push (style: fix lint/formatting), re-run.Check if the changes warrant e2e tests.
If appropriate:
docker compose ps — offer ./scripts/quick_start.sh if not runningIf changes are purely internal with no API surface change, skip with a note.
CHANGELOG.md to match formatgh pr view --json number --jq .numberdev/OBJECTIVE.mddev/NOTES.mdgit add -A
git commit -m "chore: <objective-handle> is ready"
git push
gh pr ready
This phase loops until the PR is approved and all checks pass.
gh pr checks <number>
gh pr view <number> --json reviews,comments,state,reviewDecision
If checks failed:
gh run view <run_id> --log-failedFetch comments:
gh pr view <number> --json reviews --jq '.reviews[] | select(.state != "APPROVED") | "\(.author.login) (\(.state)): \(.body)"'
gh api repos/<owner>/<repo>/pulls/<number>/comments --jq '.[] | "\(.path):\(.line // .original_line) — \(.user.login): \(.body)"'
For each comment:
fix: address review — <summary>)If approved and checks pass: tell the user the PR is ready to merge. Do NOT merge.
If changes requested or checks failing: loop back to step 11.
If waiting for review: tell the user and ask if they want to wait or come back later.
Monitor PR state. When MERGED:
If a Trello card was linked:
~/bin/trello move <CARD_ID> <DONE_LIST_ID>Ticket complete:
PR: <URL> (merged)
CHANGELOG: <entry>
Trello: <moved to Done / not linked>
Done.
Provides behavioral guidelines to reduce common LLM coding mistakes, focusing on simplicity, surgical changes, assumption surfacing, and verifiable success criteria.
Searches, retrieves, and installs Agent Skills from prompts.chat registry using MCP tools like search_skills and get_skill. Activates for finding skills, browsing catalogs, or extending Claude.
Creates, edits, and optimizes skills for Claude Code, including drafting, evaluating with test prompts, iterating on performance, and improving skill descriptions for better triggering accuracy.
npx claudepluginhub luthienresearch/internal-utils --plugin dev-pipeline