do:deepen-plan

What problem does it solve?

This Skill automates the enhancement of existing plans by orchestrating parallel research, skill, and review agents, while ensuring all outputs are persisted to disk to prevent context exhaustion.

Core Features & Use Cases

• Context-Lean Operation: All agent outputs are written to disk, keeping the parent agent's context lean and allowing for longer sessions.
• Persistent Outputs: All outputs are retained across runs for traceability and learning, with each run getting its own numbered directory.
• Automated Discovery: Discovers available skills, learnings, and review/research agents dynamically.
• Use Case: You have a complex software feature plan. Use this Skill to automatically research best practices, identify potential security vulnerabilities, and analyze the architecture, all while ensuring the detailed findings are saved for later review without filling up the AI's memory.

Quick Start

Use the do:deepen-plan skill to deepen the plan located at 'plans/feature-x-plan.md'.