agentsclimarketplace

Done

Skill endlessblink/master-plan/skills/done

Task completion workflow — run tests, commit, push, and update MASTER_PLAN.md. Use when a task is finished and ready to ship. Triggers on "/done", "mark done", "task complete", "finish task", "ship it".From its SKILL.md

Install
npx -y skills add endlessblink/master-plan --skill done

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

  • runs commandsInstructs the agent to run 8 commands, including `git status` and 7 more.

SKILL.md

4.3 KB, ~1.2k tokens by cl100k_base, as published. Nobody here has run it

Task Completion Workflow

Finalize a completed task: verify tests pass, commit changes, update MASTER_PLAN.md, and push.

Triggers

  • /master-plan:done - Main command
  • "mark done", "task complete", "finish task", "ship it"

Workflow

Step 1: Check What Changed

Run git status to see what files were modified:

git status
git diff --stat

Output immediately:

Files changed:
- [list of modified files]

Step 2: Get Task Information

Use AskUserQuestion to collect:

  1. Task ID (header: "Task ID")

    • Options: "Tracked task (enter ID)" or "Quick fix (no task ID)"
  2. Run tests? (header: "Tests")

    • Options: "Yes — run test suite (Recommended)" or "Skip tests"

Then ask in plain text: "What's a brief summary of the changes? (1-2 sentences)"

IMPORTANT: Wait for user to provide the summary before proceeding.

Step 3: Run Tests (unless skipped)

Detect and run the project's test command. Check in order:

  1. If package.json exists → npm test
  2. If Cargo.toml exists → cargo test
  3. If pyproject.toml or setup.py exists → pytest
  4. If Makefile with test target → make test
  5. If go.mod exists → go test ./...

If tests fail: STOP immediately. Report failures. Do NOT proceed. If tests pass: Continue. If no test command found: Warn user and continue.

Step 4: Update MASTER_PLAN.md (if tracked task)

Skip this step if user selected "Quick fix (no task ID)".

CRITICAL: Tasks may appear in multiple locations. Update ALL of them:

4a. Summary/Roadmap Table

Find the task row and update:

# Before:
| **TASK-XXX** | **Title** | **P2** | PLANNED | ... |

# After:
| ~~**TASK-XXX**~~ | ✅ **Title** | **P2** | ✅ DONE (YYYY-MM-DD) | ... |

4b. Subtask/Bullet Lists

# Before:
- TASK-XXX: Description

# After:
- ~~TASK-XXX~~: ✅ Description

4c. Detailed Section Header

# Before:
### TASK-XXX: Title (IN PROGRESS)

# After:
### ~~TASK-XXX~~: Title (✅ DONE)

4d. Verify All Updated

Search for the task ID and confirm all occurrences show strikethrough or ✅ DONE:

grep "TASK-XXX" docs/MASTER_PLAN.md

Step 5: Commit and Push

Stage all relevant files. NEVER commit:

  • .env* files
  • credentials*.json or files containing secrets
  • node_modules/, __pycache__/, target/, .venv/
  • Backup files, generated stats, lock files (unless intentional)

Prefer staging specific files by name over git add -A.

# Stage code files
git add <changed-files>
git add docs/MASTER_PLAN.md  # if tracked task

# Commit
git commit -m "$(cat <<'EOF'
feat(TASK-XXX): Brief summary from user

- Key change 1
- Key change 2

Co-Authored-By: Claude <[email protected]>
EOF
)"

# Push
git push

For quick fixes (no task ID):

git commit -m "$(cat <<'EOF'
fix: Brief summary from user

Co-Authored-By: Claude <[email protected]>
EOF
)"
git push

Commit prefix conventions:

  • feat(TASK-XXX): — New feature
  • fix(BUG-XXX): — Bug fix
  • chore(TASK-XXX): — Maintenance
  • refactor(TASK-XXX): — Code refactoring
  • docs(TASK-XXX): — Documentation only

Step 6: Report Completion

Output this summary:

## Task Complete

- **Task**: TASK-XXX (or "Quick fix")
- **Summary**: [user's summary]
- **Tests**: ✅ Passed (or ⏭️ Skipped)
- **Commit**: [short hash] — [message]
- **Push**: ✅ Pushed to origin
- **MASTER_PLAN.md**: Updated / N/A

Important Rules

  1. NEVER skip commit/push — Changes must be pushed to the remote
  2. ALWAYS collect summary — Don't proceed without knowing what changed
  3. Update ALL locations in MASTER_PLAN.md for tracked tasks
  4. Verify with grep after updating MASTER_PLAN.md
  5. Wait for user input — Don't assume or skip questions
  6. Test failures block completion — If tests fail, stop and report

Files to NEVER Commit

  • .env* files (secrets)
  • credentials*.json
  • Private keys, tokens
  • Generated/cached files (node_modules/, dist/, __pycache__/)
  • OS files (.DS_Store, Thumbs.db)

What ships with it

Read from the repository

Just SKILL.md. No reference files, no scripts.

Keep looking

Skills are one crate of 325,949. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.