task-decomposition

task-decomposition is a skill for Claude Code, Codex from makifbaysal/tasktrooper. It costs 33 tokens per session (2,950 once invoked), scanned A, original, Apache-2.0.

A workflow for turning an approved technical plan into separate implementation tasks. It assigns each task to a repository and code layer, then records acceptance criteria and dependencies.

In plain words
What is it for?
Use it to create implementation-board tasks for backend, frontend, or mobile work, with reviewable units, acceptance criteria, and dependency order.
Why use it?
It prevents large tasks from mixing unrelated layers or blocking parallel work unnecessarily. Explicit ordering also makes dependencies, such as an API needed by a frontend, clear to the task system.

Skill for Claude CodeCodex

Written for no agent in particular: nothing here depends on one.

Good fit Use it to create implementation-board tasks for backend, frontend, or mobile work, with reviewable units, acceptance criteria, and dependency order.

Compare 6 skills from other repositories ↓
Install with agentmods
npx agentmods add skills/makifbaysal/tasktrooper/task-decomposition
Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

Any agent
npx skills add makifbaysal/tasktrooper --skill task-decomposition
Clone the repo
git clone --depth 1 https://github.com/makifbaysal/tasktrooper

Made for: Claude Code, Codex.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for task-decomposition

README.md
[![agentmods](https://agentmods.dev/badge/skills/makifbaysal/tasktrooper/task-decomposition/github.svg)](https://agentmods.dev/skills/makifbaysal/tasktrooper/task-decomposition)
Your own site
<a href="https://agentmods.dev/skills/makifbaysal/tasktrooper/task-decomposition"><img src="https://agentmods.dev/badge/skills/makifbaysal/tasktrooper/task-decomposition/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

agentmods 80×15 button for task-decomposition

Your own site · 80×15
<a href="https://agentmods.dev/skills/makifbaysal/tasktrooper/task-decomposition"><img src="https://agentmods.dev/badge/skills/makifbaysal/tasktrooper/task-decomposition.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 33 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 2,950 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5.1 $0.00033 $0.02950
Opus 5.5 $0.00013 $0.01180
Sonnet 5.5 $0.00007 $0.00590
Haiku 4.5 $0.00003 $0.00295

Measured 4d ago against content hash cf4499e30718, method: parsed. Prices are Anthropic first-party input rates as of 2026-10-07, from the pricing page.

Security

Grade A, and why

task-decomposition scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 4d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

catalog/agents/system-architect/skills/task-decomposition/SKILL.md · 77 lines

How it starts

The opening of the file, as written. The whole thing — 77 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Task Decomposition

Overview

After the plan is written and self-reviewed, turn it into implementation board tasks with create_board_task. Decomposition quality decides whether developers can work in parallel without stepping on each other.

Slicing Rules

  • One repository + one layer per task (backend / frontend / mobile / data / game). Never bundle layers — a task that says "add the API and the UI" is two tasks, and so is "train the model and serve it from the API" when the serving endpoint lives in a backend service.
  • Each task maps to one or more plan tasks that form an independently deliverable unit: its tests can pass and its code can be reviewed without waiting for a sibling.
  • Order by dependency: backend API before the frontend/mobile that consumes it. Declare it in the ARGUMENTS (see "Ordering is an argument" below), and write the human-readable "Depends on: — consumes POST /api/v1/..." line in the description as well — but never only the line, because prose enforces nothing.
  • A project with UI and no design system: call get_design_system (with the project's project_id) before slicing UI work. If it returns no base — and the UI repository no layer — never hand-write a design task for it: call request_design_system (scope: "project" with the project_id; scope: "repository" with its repository_id for a UI repository in no project), with what the analysis knows about the audience, the feel and the brief's references in notes. It opens the same design task the project's Design System tab opens — the ui-designer assigned, the server's brief as its description — or returns the one already open (created: false); its returned key is what every UI task and every screen design task below waits for with blocked_by (it lives in one UI repository; a blocker in another repository is fine). On approval the designer opens an "Apply design system vN" task per repository that writes the generated design files and builds the foundation (tokens, atomic folders + INVENTORY.md, base atoms, ui-guard) from them — so do not also create a separate "Add design system foundation" task. The UI tasks in one repository are chained with blocked_by (shared files serialize, below); the first one writes the generated files from get_design_system files: true verbatim before its screen — the same bytes the Apply task writes.
  • Design first for a screen nobody has designed: a frontend or mobile task whose screen has no approved design (no design task already covers it) and no matching existing screen to follow gets a screen design task first — create_board_task with task_type: "design", no assignee, repository (the repository the screen ships in), derived_from this analiz task, a description naming the screen(s), the audience and the states the plan needs, and blocked_by the design-system task when you requested one above (a screen is drawn only with an approved design system). The frontend/mobile tasks for that screen are blocked_by the screen design task — that alone hands their runs its approved mockups and hand-off spec, and they start only once the designer has written the hand-off and released it; their derived_from stays this analiz task. A change that follows an existing screen's pattern needs no design task.
  • Size: a task should land at an estimated ≤ ~400 changed lines including tests, so a reviewer sees the whole diff inside the 24,000-byte cut (review-reads-whole-diff); over ~1,000 is a split. Keep refactors in their own task, separate from feature work.
  • Shared files serialize: two tasks in one repository that both add a migration (sequence numbers collide), or edit the same router table, locale bundle, DI container, INVENTORY.md or OpenAPI file, are chained with blocked_by rather than left to collide at review time.

Read the full file on GitHub · 77 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 4d ago Changed · +2 lines cf4499e30718
  2. 8d ago First seen · 75 lines · 33 tokens per session scan A be06c5c849da

Subscribe to this mod's changes

task-decomposition is a skill published in the GitHub repository makifbaysal/tasktrooper (113 stars, last pushed yesterday), licensed Apache-2.0. It adds 33 tokens to every session and 2,950 once invoked, about $0.0001 per session on Opus 5.5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-10-05.

Related

Other skills, from other repositories

github-multi-repo

Multi-repository coordination, synchronization, and architecture management with AI swarm orchestration.

ruvnet/ruflo · 21 tokens

paperclip

Use for Paperclip-managed tasks and heartbeats: reading task context, delivering task documents or files, updating completion or blockers, coordinating or delegating work, and following company governance. Includes control plane API operations for assignments, comments, approvals, and routines.

paperclipai/paperclip · 54 tokens

paperclip

Interact with the Paperclip control plane API for task coordination and governance. Use when checking assignments, updating issue status, posting comments, delegating work, managing routines, or calling Paperclip API endpoints.

paperclipai/paperclip · 43 tokens

paperclip-board

Manage a Paperclip company as a board member via chat. Use when the user wants onboarding, company or agent management, approvals, task monitoring, cost oversight, or work product review in the Paperclip control plane.

paperclipai/paperclip · 47 tokens

paperclip-distill

Use when an operation issue is a Paperclip cursor-window, distill, or backfill. Turn source-bundled Paperclip activity into wiki-insightful project standups, durable project pages, decisions, and history without asset dereferencing.

paperclipai/paperclip · 54 tokens

summarize-status

Write a short, colloquial summary for a Paperclip summary slot: open with the 1–3 specific, concrete actions the reader needs to take right now to unblock the work, then a brief plain-language status, streaming progress as it works.

paperclipai/paperclip · 0 tokens