$npx -y skills add DevelopersGlobal/ai-agent-skills --skill task-decompositionBreaks features into atomic, independently verifiable tasks. No task should take more than 4 hours. Unblocks parallel work and reduces integration risk.
| 1 | ## Overview |
| 2 | |
| 3 | Large tasks fail in large ways. Decomposed tasks fail in small, fixable ways. This skill breaks any feature or project into atomic tasks — each independently implementable, testable, and deployable. |
| 4 | |
| 5 | ## When to Use |
| 6 | |
| 7 | - Before starting any feature that takes more than half a day |
| 8 | - When a task feels overwhelming or unclear |
| 9 | - When multiple people need to work in parallel |
| 10 | |
| 11 | ## Process |
| 12 | |
| 13 | ### Step 1: Identify the Deliverable |
| 14 | |
| 15 | 1. State what "done" looks like for the whole feature. |
| 16 | 2. List all the things that must be true when it's complete. |
| 17 | 3. Identify dependencies: what must exist before any task can start? |
| 18 | |
| 19 | **Verify:** You can state the full feature goal in 2 sentences. |
| 20 | |
| 21 | ### Step 2: Decompose Into Atomic Tasks |
| 22 | |
| 23 | 4. Break the feature into tasks where each task: |
| 24 | - Can be completed in under 4 hours |
| 25 | - Has a single, clear output |
| 26 | - Can be verified independently |
| 27 | - Can be reverted without breaking other tasks |
| 28 | 5. Each task should be: `[verb] [noun] so that [outcome]` |
| 29 | - ✅ "Add rate limiting to /api/login so that brute-force is prevented" |
| 30 | - ❌ "Work on the login security stuff" |
| 31 | |
| 32 | **Verify:** Every task is under 4 hours. Every task has a clear verify condition. |
| 33 | |
| 34 | ### Step 3: Order and Parallelism |
| 35 | |
| 36 | 6. Draw the dependency graph — what blocks what? |
| 37 | 7. Identify tasks that can be done in parallel. |
| 38 | 8. Sequence tasks so integration happens incrementally (not as one big bang at the end). |
| 39 | |
| 40 | **Verify:** The dependency order is clear. No unnecessary sequential dependencies. |
| 41 | |
| 42 | ### Step 4: Estimate and Adjust |
| 43 | |
| 44 | 9. For each task, estimate: best case / worst case / expected. |
| 45 | 10. If any task's worst case > 1 day: decompose further. |
| 46 | 11. Total estimate sanity check: does it add up to a reasonable timeline? |
| 47 | |
| 48 | ## Common Rationalizations (and Rebuttals) |
| 49 | |
| 50 | | Excuse | Rebuttal | |
| 51 | |--------|----------| |
| 52 | | "I'll figure it out as I go" | Decomposition takes 30 minutes. Rework from poor planning takes days. | |
| 53 | | "It's too complex to break down" | Everything can be decomposed. Start with what you know, decompose the rest later. | |
| 54 | | "The tasks are too granular" | Granular tasks ship continuously. Coarse tasks ship never. | |
| 55 | |
| 56 | ## Verification |
| 57 | |
| 58 | - [ ] Each task is under 4 hours |
| 59 | - [ ] Each task has a single verify condition |
| 60 | - [ ] Dependencies mapped — no circular dependencies |
| 61 | - [ ] Parallel work identified |
| 62 | |
| 63 | ## References |
| 64 | |
| 65 | - [goal-driven-execution skill](../goal-driven-execution/SKILL.md) |
| 66 | - [multi-agent-orchestration skill](../multi-agent-orchestration/SKILL.md) |