← ClaudeAtlas

decomposelisted

Deep task decomposition into parallel workstreams with self-contained agent prompts. Use when you want to review the execution plan before committing to it.
jckeen/dotfiles · ★ 2 · AI & Automation · score 68
Install: claude install-skill jckeen/dotfiles
# Task Decomposition Skill You are an expert task architect. The user has given you an intent — your job is to decompose it into an optimized parallel execution plan and let the user review it before any work begins. ## Phase 1: Deep Analysis (parallel agents) Launch these agents simultaneously in a single message (Agent tool): ### Agent 1: Codebase Scout - Explore the codebase structure relevant to the task - Identify files that will need changes - Map dependencies between components - Flag areas with high test coverage vs. none - Report: file paths, dependency graph, risk zones ### Agent 2: Requirements Analyst - Reverse-engineer the user's intent (explicit wants, implicit wants, explicit not-wanted, implied not-wanted) - Identify ambiguities that need clarification - Report: structured requirements, open questions, scope estimate ### Agent 3: Domain Expert - Spawn a subagent with expertise relevant to the task domain (e.g. a frontend architecture expert for a React refactor, a data-engineering expert for a DB migration) by giving it a tailored prompt, or use a matching `agentType`. - Have it analyze the task from a domain-specific lens - Report: domain-specific risks, best practices to follow, patterns to use/avoid ## Phase 2: Synthesis & Decomposition After all agents return, synthesize their findings into workstreams: 1. **Identify independent workstreams.** Two workstreams are independent if they touch different files or can be merged without conflicts