From dev-pipeline
Ticket kickoff only (setup + plan + draft PR + failing tests). Use /dev for the full autonomous pipeline. Accepts goal or Trello card. Example: /start-ticket implement rate limiting
How this skill is triggered — by the user, by Claude, or both
Slash command
/dev-pipeline:start-ticketThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Kicks off a new piece of work — from raw goal to failing tests that define "done." Use this when you want to set up and plan but implement manually. For full automation, use `/dev` instead.
Kicks off a new piece of work — from raw goal to failing tests that define "done." Use this when you want to set up and plan but implement manually. For full automation, use /dev instead.
Commit early and often. Push after every commit. Don't accumulate local commits.
The user provides one of:
/start-ticket implement rate limiting for the admin API/start-ticket https://trello.com/c/abc123If no argument, ask: "What are you working on? (paste a goal or Trello card URL)"
Run pwd. Must contain .claude/worktrees/.
If not: STOP. Say: "Not in a worktree. Run claude --worktree first, then /start-ticket again."
If .env doesn't exist:
cp ../../.env .env (from main project dir)cp .env.example .envTrello card (URL contains trello.com/c/ or input starts with trello:):
~/bin/trello card <ID>~/bin/trello move <ID> <IN_PROGRESS_LIST_ID> (look up list ID via ~/bin/trello lists <BOARD_ID> if needed)/finish-ticketPlain text: use as-is.
Before planning, understand the relevant code:
dev/context/ files for relevant learnings, decisions, gotchas# Objective
<One-sentence goal>
## Description
<2-3 sentences of context, informed by codebase research>
## Approach
<Bulleted plan — which files to modify, what the change looks like>
## Acceptance Criteria
- [ ] Failing unit tests written and verified failing
- [ ] Implementation makes all tests pass
- [ ] dev_checks passes
- [ ] <domain-specific criteria from the goal>
## Tracking
- Trello: <card URL or "none">
- Branch: <branch name>
- PR: <filled after creation>
Present the approach and ask: "Does this plan look right?"
Wait for confirmation before proceeding.
Get the draft PR up before writing any tests or code:
git add dev/OBJECTIVE.md
git commit -m "chore: set objective to <short-handle>"
git push -u origin $(git rev-parse --abbrev-ref HEAD)
gh pr create --draft --title "<type>: <title>" --body "$(cat <<'EOF'
## Objective
<objective statement>
## Approach
<approach from OBJECTIVE.md>
## Acceptance Criteria
<criteria>
EOF
)"
Update dev/OBJECTIVE.md Tracking section with the PR URL. If Trello card is linked, comment the PR URL on the card.
Write unit tests that define the expected behavior:
Run the new tests: uv run pytest <test_file> -v
Confirm they fail because the feature doesn't exist, not due to infrastructure issues.
git add <test_files>
git commit -m "test: add failing tests for <feature>"
git push
Ready to implement:
Worktree: <path>
Branch: <branch>
PR (draft): <URL>
Trello: <card URL — moved to In Progress / not linked>
Failing tests: <N tests in M files>
The tests define what "done" looks like. Implement until they pass.
npx claudepluginhub luthienresearch/internal-utils --plugin dev-pipelineProvides 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.