Claude Code Plugins

Community-maintained marketplace

Feedback

executing-tickets

@jssee/.dotfiles
0
0

Use to implement tickets created by writing-tickets skill

Install Skill

1Download skill
2Enable skills in Claude

Open claude.ai/settings/capabilities and find the "Skills" section

3Upload to Claude

Click "Upload skill" and select the downloaded ZIP file

Note: Please verify skill by going through its instructions before using it.

SKILL.md

name executing-tickets
description Use to implement tickets created by writing-tickets skill

Executing Tickets

Overview

Work through tickets from tk ready, implementing each one following TDD. Each ticket is a self-contained unit of work with all context needed.

Announce at start: "I'm using the executing-tickets skill to implement tickets."

Context: Run in the worktree where tickets were created.

Workflow

1. Get Next Ticket

tk ready

Pick the highest priority ticket (or ask user if multiple options).

2. Start Work

tk start <id>
tk show <id>

Read the ticket thoroughly. Understand:

  • Files to create/modify
  • Steps to follow
  • Acceptance criteria

3. Execute Steps

Follow the ticket's steps exactly. Typical TDD flow:

  1. Write failing test - Copy test code from ticket
  2. Run test - Verify it fails as expected
  3. Implement - Write minimal code to pass
  4. Run test - Verify it passes
  5. Commit - Use commit message from ticket or derive from task

4. Verify Acceptance Criteria

Before closing, check every acceptance criterion:

  • All tests pass
  • Code follows project patterns
  • No regressions (run full test suite if ticket impacts shared code)

5. Close Ticket

tk close <id>

Add notes if needed:

tk add-note <id> "Deviated from plan: used X instead of Y because..."
tk add-note <id> "Found issue: created follow-up ticket for Z"

6. Repeat

tk ready

Continue until no tickets ready or user stops.

Handling Blockers

If stuck on a ticket:

# Check what's blocking
tk blocked

# If missing dependency, work on blocker first
tk show <blocker-id>

# If unclear requirements, ask user
# If bug in plan, add note and ask user
tk add-note <id> "Blocked: [reason]"

Creating Follow-up Tickets

If you discover additional work needed:

tk create "Fix edge case discovered in X" --type=bug --priority=1
tk dep <new-id> <current-epic-id>
tk add-note <current-id> "Created follow-up: <new-id>"

Code Review Checkpoints

After each ticket:

  1. Review diff: git diff HEAD~1
  2. Run linter/formatter
  3. Run relevant tests
  4. Sanity check: does this actually solve the ticket?

If issues found, fix before moving to next ticket.

Commands Reference

tk ready                 # next workable tickets
tk start <id>            # mark in_progress
tk show <id>             # view ticket details
tk close <id>            # mark done
tk add-note <id> "..."   # append note
tk blocked               # view blocked tickets
tk dep tree <epic-id>    # see overall progress
tk ls --status=in_progress  # what's currently active

Remember

  • One ticket at a time
  • Follow steps exactly (they contain tested code)
  • Commit after each ticket
  • Don't skip acceptance criteria
  • Add notes for deviations or discoveries
  • Ask user if unclear, don't guess