decomposing-plans

Decompose monolithic implementation plans into task files with dependency analysis and parallel batch scheduling.

3|Updated Nov 19, 2025
One-click install
npx skills add https://github.com/seanGSISG/crispy-claude --skill decomposing-plans
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: decomposing-plans
Source: https://github.com/seanGSISG/crispy-claude/tree/main/.claude/skills/decomposing-plans
Command: npx skills add https://github.com/seanGSISG/crispy-claude --skill decomposing-plans

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) components.

What problem does it solve?

Monolithic implementation plans can be inefficient for AI agents, leading to high token usage and sequential execution. This skill automates the decomposition of large plans into smaller, parallelizable tasks.

Core Features & Use Cases

  • Automated Task Extraction: Parses monolithic plans to extract individual tasks, saving context tokens for subagents.
  • Dependency Analysis: Identifies explicit and file-based dependencies between tasks to ensure correct execution order.
  • Parallel Batch Identification: Groups tasks into parallel batches (up to 2 concurrent subagents) for faster execution.

Quick Start

Use the decomposing-plans skill to break down the 'user-authentication' plan from 'docs/plans/2025-01-18-user-auth.md' into parallel tasks.

Frequently Asked Questions about decomposing-plans

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I break down a large implementation plan into parallel tasks?

Decomposing plans automatically extracts individual tasks from monolithic markdown files, analyzes dependencies, and groups tasks into parallel batches (up to 2 concurrent subagents). This generates separate task files and a manifest for efficient scheduling and execution.

What's the best way to reduce token usage when executing long implementation plans with AI agents?

Breaking monolithic plans into smaller, parallelizable tasks significantly reduces token usage per subagent. Task decomposition enables parallel execution across multiple agents instead of sequential processing of one large document.

How do I identify which tasks in my plan can run in parallel?

Decomposing plans analyzes explicit and file-based dependencies between tasks to determine which ones have no blocking relationships. Tasks without dependencies on each other are grouped into parallel batches for concurrent execution.

Can I automate the creation of individual task files from a single implementation plan document?

Yes. Decomposing plans parses markdown plan files, extracts task-level content and metadata, and automatically generates per-task files organized in a task directory structure with a manifest.json for scheduling.

What file formats does plan decomposition work with?

Decomposing plans reads monolithic implementation plans stored as markdown files (docs/plans/YYYY-MM-DD-<feature-name>.md) and outputs individual task markdown files plus a JSON manifest for execution.

Do I need to manually specify task dependencies, or does decomposition detect them automatically?

Decomposing plans automatically identifies dependencies by analyzing the monolithic plan structure and tracking which files each task modifies. Manual dependency specification is not required.