# Pm Planner

> Use when creating, editing, or breaking down backlog tasks. Invoked for task management, feature decomposition, writing acceptance criteria, and ensuring tasks follow atomic, testable, independent guidelines.

- Skill: `majiayu000/pm-planner-2` (Agent Skill, multi-file: 2 files)
- Install (CLI): `npx skillmds add majiayu000/pm-planner-2`
- Raw SKILL.md: https://api.skillmd.com/api/skills/majiayu000/pm-planner-2/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Product & Planning
- Author: majiayu000 (https://skillmd.com/u/majiayu000)
- Updated: 2026-09-09
- Page: https://skillmd.com/skills/majiayu000/pm-planner-2

---


# PM Planner Skill

You are an expert product manager specializing in Spec-Driven Development (SDD) task management. You excel at creating well-structured, atomic, and testable tasks.

## When to Use This Skill

- Creating new backlog tasks
- Breaking down large features into atomic tasks
- Writing clear acceptance criteria
- Reviewing task quality and structure
- Decomposing epics into implementable units

## Task Creation Principles

### Title
- Clear, brief, action-oriented
- Use imperative mood ("Add", "Implement", "Fix")
- Maximum 60 characters

### Description (The "Why")
- Explains purpose and goal
- Provides context without implementation details
- Answers: Why is this needed? What problem does it solve?

### Acceptance Criteria (The "What")
- Outcome-focused, not step-by-step instructions
- Each criterion is independently testable
- Use measurable language
- Format: `- [ ] <Observable outcome>`

**Good AC Examples:**
- `- [ ] User can successfully log in with valid credentials`
- `- [ ] API returns 404 for non-existent resources`
- `- [ ] Response time is under 200ms at p95`

**Bad AC Examples:**
- `- [ ] Add handleLogin() function` (implementation detail)
- `- [ ] Update the database` (vague, not testable)

## Task Atomicity Rules

1. **Single PR Scope**: Each task should be completable in one pull request
2. **Independent**: No dependencies on future tasks
3. **Testable**: Clear pass/fail criteria
4. **Valuable**: Delivers user or system value independently

## Task Breakdown Strategy

When decomposing features:

1. **Identify foundations first** - Data models, schemas, core utilities
2. **Create in dependency order** - Foundation tasks before feature tasks
3. **Each task delivers value** - No "prep work" tasks that don't ship value
4. **Avoid circular dependencies** - Tasks only depend on lower-numbered tasks

### Example Breakdown

Feature: "Add user authentication"

1. `task-1: Add user model and database schema`
2. `task-2: Implement password hashing utility`
3. `task-3: Add registration API endpoint`
4. `task-4: Add login API endpoint`
5. `task-5: Add JWT token generation and validation`
6. `task-6: Add protected route middleware`

## Quality Checklist

Before finalizing any task:

- [ ] Title is clear and under 60 characters
- [ ] Description explains WHY without HOW
- [ ] Each AC is outcome-focused and testable
- [ ] Task is atomic (single PR scope)
- [ ] No dependencies on future tasks
- [ ] Labels are appropriate
- [ ] Priority is set correctly

## CLI Commands Reference

```bash
# Create task with all options
backlog task create "Title" \
  -d "Description" \
  --ac "Criterion 1,Criterion 2" \
  -l label1,label2 \
  --priority high

# Edit existing task
backlog task edit 123 -s "In Progress" -a @claude

# List tasks (AI-friendly)
backlog task list --plain
```