Task Decomposition
Split a feature or engineering task into independently shippable, verifiable subtasks with explicit dependency ordering. Use when the task is too big for a single implementation pass, when you need to parallelize work across multiple agents, or when you need confidence in a phased deliverable.
MCP get_skill({ skillId: "task-decomposition" })Use this skill with your agent
Create a free account and connect via MCP
--- title: "Task Decomposition" description: "Split a feature or engineering task into independently shippable, verifiable subtasks with explicit dependency ordering. Use when the task is too big for a single implementation pass, when you need to parallelize work across multiple agents, or when you need confidence in a phased deliverable." author: AgentArmory license: Proprietary --- # Task Decomposition Split a feature into independently shippable, verifiable subtasks. This is the skill you load BEFORE starting Implementation Workflow, when the feature is complex enough that a single implementation pass would risk failure cascades, missing dependencies, or unbounded scope. Research shows that AI agents working on long-horizon tasks degrade in code quality with each turn (SlopCodeBench, 2026) and that the #1 cited improvement for AI coding is "better task decomposition" (Stack Overflow, 2026). This skill is designed to contain that degradation. ## Table of Contents - [When to Use](#when-to-use) - [Triggers](#triggers) - [Prerequisites](#prerequisites) - [Methodology](#methodology) - [Dos](#don'ts) - [Dos](#dos) - [Pitfalls](#pitfalls) - [Nonsense Check](#nonsense-check) - [Validate Before Shipping](#validate-before-shipping) - [Cross-Skill Hints](#cross-skill-hints) ## When to Use Use when: - The task will require more than ~15 minutes of agent execution time (a single session) - The feature touches multiple files, components, APIs, or databases - You need to parallelize work across multiple agents - The task is too large to verify in one pass - You are planning a multi-PR feature and need a release sequence - The user asks "what needs to happen first?" Do NOT use for: a single function or simple change (use Implementation Workflow directly), debugging (use Systematic Debugging), or a question / research task. ## Triggers "break this down", "subtasks", "decompose", "split into", "what are the steps", "parallelize", "phases", "milestones", "what's the order", "dependency ordering", "cut this into pieces", "chunk this" ## Prerequisites - A clear understanding of the feature goal (acceptance criteria, spec, or ticket) - Knowledge of the codebase structure and key files (if code-level decomposition is needed) - The project's deployment / release cadence (affects what constitutes a shippable unit) - Implementation Workflow skill loaded (this is a planning prelude to it) ## Methodology
Related Skills
More skills in Software Engineering
Accessibility Standards
Comprehensive web accessibility standards based on WCAG 2.2 AA, with 38+ anti-patterns, legal enforcement context (EAA, ADA Title II), WAI-ARIA patterns, and framework-specific fixes for modern web frameworks and libraries.
Accord
Authoring unified specification packages across Business/Development/Design teams via staged elaboration (L0 Vision → L1 Requirements → L2 Team Detail → L3 Acceptance Criteria). No code. Use when authoring cross-team specs, building L0-L3 packages, or aligning Biz/Dev/Design on a single source of truth.
Acquire Codebase Knowledge
Use this skill when the user explicitly asks to map, document, or onboard into an existing codebase. Trigger for prompts like "map this codebase", "document this architecture", "onboard me to this repo", or "create codebase docs". Do not trigger for routine feature implementation, bug fixes, or narrow code edits unless the user asks for repository-level discovery.
Acreadiness Assess
Run the AgentRC readiness assessment on the current repository and produce a static HTML dashboard at reports/index.html. Wraps `npx github:microsoft/agentrc readiness` and hands off rendering to the @ai-readiness-reporter custom agent. Supports policies (--policy) for org-specific scoring. Use when asked to assess, audit, or score the AI readiness of a repo.
Acreadiness Generate Instructions
Generate tailored AI agent instruction files via AgentRC instructions command. Produces .github/copilot-instructions.md (default, recommended for Copilot in VS Code) plus optional per-area .instructions.md files with applyTo globs for monorepos. Use after running /acreadiness-assess to close gaps in the AI Tooling pillar.
Acreadiness Policy
Help the user pick, write, or apply an AgentRC policy. Policies customise readiness scoring by disabling irrelevant checks, overriding impact/level, setting pass-rate thresholds, or chaining org baselines with team overrides. Use when the user asks about strict mode, AI-only scoring, custom weights, CI gating, or wants org-wide standardisation.