engineer-work

Execute implementation plans with git branching, progress tracking, quality checks, and PR creation. Use when the user says "build this", "implement this", "start working", "execute the plan", or provides a plan file to execute.

Safety Notice

This listing is imported from skills.sh public index metadata. Review upstream SKILL.md and repository scripts before running.

Copy this and send it to your AI assistant to learn

Install skill "engineer-work" with this command: npx skills add elliottrjacobs/bench-skills/elliottrjacobs-bench-skills-engineer-work

/engineer-work — Plan Execution

Execute implementation plans task-by-task with structured progress tracking, incremental commits, quality checks, and PR creation.

When to Use

  • User says "build this", "implement this", "start working"
  • After completing /engineer-plan and ready to execute
  • User provides a plan file to implement

Process

Step 1: Setup

Locate the plan: Use $ARGUMENTS as the plan path, or check docs/plans/ for the most recent plan, or ask the user.

Create a branch:

git checkout -b feature/<plan-name>

Initialize tracking: Create a TodoWrite with all tasks from the plan. Mark Task 0 (setup) as in_progress.

Step 2: Execute Tasks

Work through each task in dependency order.

Per-task loop:

  1. Mark task in_progress in TodoWrite
  2. Read the task — review acceptance criteria, key files
  3. Implement — follow existing codebase patterns, keep changes minimal
  4. Self-check — verify acceptance criteria are met
  5. Commit — one commit per logical unit of work
  6. Mark task completed in TodoWrite

Implementation rules:

  • Follow existing patterns (don't introduce new conventions)
  • Keep changes minimal — implement what the task says, nothing more
  • Run npx tsc --noEmit after significant changes
  • Fix lint errors as you go
  • If stuck, check docs/solutions/ for relevant past solutions
  • If a task reveals a plan problem, note it and adapt

Commit format:

type(scope): description

Types: feat, fix, refactor, chore, test, docs

  • Imperative mood: "add" not "added"
  • Under 72 characters
  • One commit per task (split large tasks into multiple)

Step 3: Quality Checks

After all tasks complete, detect the tech stack and run appropriate checks.

Always run:

npx tsc --noEmit          # TypeScript type check
npm run lint              # or project-specific lint command

Stack-adaptive checks (detect from config files):

  • Next.js (next.config.*): npm run build to catch build errors
  • Expo (app.json): npx expo lint
  • Tests: npm test if test scripts exist

Fix any issues found, commit as fix: address lint/type issues, and re-run checks.

Step 4: Create PR

gh pr create --title "[type]: [feature name]" --body "$(cat <<'EOF'
## Summary
- [Primary change]
- [Secondary change]

## Plan Reference
[Link to docs/plans/ file]

## Testing
- [ ] Type check passes
- [ ] Lint passes
- [ ] [Feature-specific verification]
EOF
)"

Step 5: Handoff

Use AskUserQuestion:

  • "PR is ready. What's next?" (header: "Next step")
    • "Run a full review (/engineer-review)" — Comprehensive code review
    • "Document learnings (/knowledge-compound)" — Capture patterns
    • "Done" — Close out

Output

  • Feature branch with incremental commits
  • Pull request on GitHub

Next Steps

  • Want a thorough review? → /engineer-review
  • Security check needed? → /security-audit
  • Capture what you learned? → /knowledge-compound

Source Transparency

This detail page is rendered from real SKILL.md content. Trust labels are metadata-based hints, not a safety guarantee.

Related Skills

Related by shared tags or category signals.

General

product-prd

No summary provided by upstream source.

Repository SourceNeeds Review
General

product-tech-spec

No summary provided by upstream source.

Repository SourceNeeds Review
General

product-brainstorm

No summary provided by upstream source.

Repository SourceNeeds Review
General

product-naming

No summary provided by upstream source.

Repository SourceNeeds Review