$npx -y skills add One-Man-Company/Skills-ContextManager --skill plan-writingStructured task planning with clear breakdowns, dependencies, and verification criteria. Use when implementing features, refactoring, or any multi-step work.
| 1 | # Plan Writing |
| 2 | |
| 3 | > Source: obra/superpowers |
| 4 | |
| 5 | ## Overview |
| 6 | This skill provides a framework for breaking down work into clear, actionable tasks with verification criteria. |
| 7 | |
| 8 | ## Task Breakdown Principles |
| 9 | |
| 10 | ### 1. Small, Focused Tasks |
| 11 | - Each task should take 2-5 minutes |
| 12 | - One clear outcome per task |
| 13 | - Independently verifiable |
| 14 | |
| 15 | ### 2. Clear Verification |
| 16 | - How do you know it's done? |
| 17 | - What can you check/test? |
| 18 | - What's the expected output? |
| 19 | |
| 20 | ### 3. Logical Ordering |
| 21 | - Dependencies identified |
| 22 | - Parallel work where possible |
| 23 | - Critical path highlighted |
| 24 | - **Phase X: Verification is always LAST** |
| 25 | |
| 26 | ### 4. Dynamic Naming in Project Root |
| 27 | - Plan files are saved as `{task-slug}.md` in the PROJECT ROOT |
| 28 | - Name derived from task (e.g., "add auth" → `auth-feature.md`) |
| 29 | - **NEVER** inside `.claude/`, `docs/`, or temp folders |
| 30 | |
| 31 | ## Planning Principles (NOT Templates!) |
| 32 | |
| 33 | > 🔴 **NO fixed templates. Each plan is UNIQUE to the task.** |
| 34 | |
| 35 | ### Principle 1: Keep It SHORT |
| 36 | |
| 37 | | ❌ Wrong | ✅ Right | |
| 38 | |----------|----------| |
| 39 | | 50 tasks with sub-sub-tasks | 5-10 clear tasks max | |
| 40 | | Every micro-step listed | Only actionable items | |
| 41 | | Verbose descriptions | One-line per task | |
| 42 | |
| 43 | > **Rule:** If plan is longer than 1 page, it's too long. Simplify. |
| 44 | |
| 45 | --- |
| 46 | |
| 47 | ### Principle 2: Be SPECIFIC, Not Generic |
| 48 | |
| 49 | | ❌ Wrong | ✅ Right | |
| 50 | |----------|----------| |
| 51 | | "Set up project" | "Run `npx create-next-app`" | |
| 52 | | "Add authentication" | "Install next-auth, create `/api/auth/[...nextauth].ts`" | |
| 53 | | "Style the UI" | "Add Tailwind classes to `Header.tsx`" | |
| 54 | |
| 55 | > **Rule:** Each task should have a clear, verifiable outcome. |
| 56 | |
| 57 | --- |
| 58 | |
| 59 | ### Principle 3: Dynamic Content Based on Project Type |
| 60 | |
| 61 | **For NEW PROJECT:** |
| 62 | - What tech stack? (decide first) |
| 63 | - What's the MVP? (minimal features) |
| 64 | - What's the file structure? |
| 65 | |
| 66 | **For FEATURE ADDITION:** |
| 67 | - Which files are affected? |
| 68 | - What dependencies needed? |
| 69 | - How to verify it works? |
| 70 | |
| 71 | **For BUG FIX:** |
| 72 | - What's the root cause? |
| 73 | - What file/line to change? |
| 74 | - How to test the fix? |
| 75 | |
| 76 | --- |
| 77 | |
| 78 | ### Principle 4: Scripts Are Project-Specific |
| 79 | |
| 80 | > 🔴 **DO NOT copy-paste script commands. Choose based on project type.** |
| 81 | |
| 82 | | Project Type | Relevant Scripts | |
| 83 | |--------------|------------------| |
| 84 | | Frontend/React | `ux_audit.py`, `accessibility_checker.py` | |
| 85 | | Backend/API | `api_validator.py`, `security_scan.py` | |
| 86 | | Mobile | `mobile_audit.py` | |
| 87 | | Database | `schema_validator.py` | |
| 88 | | Full-stack | Mix of above based on what you touched | |
| 89 | |
| 90 | **Wrong:** Adding all scripts to every plan |
| 91 | **Right:** Only scripts relevant to THIS task |
| 92 | |
| 93 | --- |
| 94 | |
| 95 | ### Principle 5: Verification is Simple |
| 96 | |
| 97 | | ❌ Wrong | ✅ Right | |
| 98 | |----------|----------| |
| 99 | | "Verify the component works correctly" | "Run `npm run dev`, click button, see toast" | |
| 100 | | "Test the API" | "curl localhost:3000/api/users returns 200" | |
| 101 | | "Check styles" | "Open browser, verify dark mode toggle works" | |
| 102 | |
| 103 | --- |
| 104 | |
| 105 | ## Plan Structure (Flexible, Not Fixed!) |
| 106 | |
| 107 | ``` |
| 108 | # [Task Name] |
| 109 | |
| 110 | ## Goal |
| 111 | One sentence: What are we building/fixing? |
| 112 | |
| 113 | ## Tasks |
| 114 | - [ ] Task 1: [Specific action] → Verify: [How to check] |
| 115 | - [ ] Task 2: [Specific action] → Verify: [How to check] |
| 116 | - [ ] Task 3: [Specific action] → Verify: [How to check] |
| 117 | |
| 118 | ## Done When |
| 119 | - [ ] [Main success criteria] |
| 120 | ``` |
| 121 | |
| 122 | > **That's it.** No phases, no sub-sections unless truly needed. |
| 123 | > Keep it minimal. Add complexity only when required. |
| 124 | |
| 125 | ## Notes |
| 126 | [Any important considerations] |
| 127 | ``` |
| 128 | |
| 129 | --- |
| 130 | |
| 131 | ## Best Practices (Quick Reference) |
| 132 | |
| 133 | 1. **Start with goal** - What are we building/fixing? |
| 134 | 2. **Max 10 tasks** - If more, break into multiple plans |
| 135 | 3. **Each task verifiable** - Clear "done" criteria |
| 136 | 4. **Project-specific** - No copy-paste templates |
| 137 | 5. **Update as you go** - Mark `[x]` when complete |
| 138 | |
| 139 | --- |
| 140 | |
| 141 | ## When to Use |
| 142 | |
| 143 | - New project from scratch |
| 144 | - Adding a feature |
| 145 | - Fixing a bug (if complex) |
| 146 | - Refactoring multiple files |