- Home
- Skills
- Productivity
- Decompose: Turn Tasks into Checkable Plans
Works with the AI Agents you already use
Decompose: Turn Tasks into Checkable Plans
Turn a complex request into checkable units, dependency waves, and a plan your agent can review.
$5
Decompose: Turn Tasks into Checkable Plans
Example session with this skill installed
Create a synthetic plan to compare two sample formats and write a recommendation. Check the unit map and show its dependency waves.
- Read your context and instructions
- Compiled the decompose
The shipped checker accepted the three-unit map. It placed the two research units in wave 0 and the recommendation in wave 1, reported coverage for both clauses, and calculated a seven-minute critical path from the supplied estimates. The advisor review was not required. See request, result, and tool output.
Connects securely to your tools. The creator never sees your data.
About this skill
Decompose helps an agent split a request into units with inputs, outputs, and checks. It validates dependencies and clause coverage, groups ready work into waves, and prints a plan table and critical path. For larger or highly connected maps, it calls for an independent advisor review before model and machine picks. The agent writes the plan and makes those picks. Decompose plans work; it does not execute the planned steps.
For simpler work, validate a step plan or expand a task card using bundled recipes. The checker catches dependency cycles and overlapping write paths. You can save validated steps as a reusable recipe.
The same skill is open source under MIT at https://github.com/aunysillyme-dev/aunx-decompose. This pack is the ready-to-drop skill folder. Place decompose in your agent's skills/ folder, then follow the skill's command instructions.
Credits
Decompose is original work by Auny LLC, MIT licensed. It borrows ideas, not code. Each technique below is credited to its source.
| Technique | Source | Where it is used |
|---|---|---|
| Topological layering | Arthur B. Kahn, "Topological sorting of large networks", Communications of the ACM 5(11), 1962 | Building parallel waves from dependencies and detecting cycles. |
| Critical path method | James E. Kelley Jr. and Morgan R. Walker, "Critical-path planning and scheduling", Eastern Joint Computer Conference, 1959 | The longest dependency chain printed with each plan. |
| Work and span bound | Richard P. Brent, "The parallel evaluation of general arithmetic expressions", Journal of the ACM 21(2), 1974 | The illustrative speedup bound T1 / Tinf in the picking reference. |
| Lead-and-workers and parallelization patterns | Anthropic, "Building effective agents", 2024 | Splitting work across parallel workers and having one lead agent merge and check the results. |
No third-party code is bundled.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Productivity.
Decompose plans work; it does not execute the planned steps. It needs Python 3.8 or later and an agent that can read SKILL.md and run Python.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before buying a skill.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed today
- Software
Python 3.8 or later
Node.jsaunx-decompose
Install as skills/decompose in your agent's skills folder.
Permissions
The helper reads the input files supplied to it and bundled recipes. It makes no network requests and starts no other programs. Saving a recipe writes to `DECOMPOSE_RECIPES` or `~/.claude/decompose/recipes.json`, with a lock and temporary file beside the store. The agent creates temporary task files as directed by the skill. Agent tools used for later work have their own permissions.