generate

Stress-test existing CLAUDE.md files with fresh agents to collect missing repo guidance.

Updated Feb 23, 2026
One-click install
npx skills add https://github.com/gonzaloetjo/setup-claude-md --skill generate-gonzaloetjo
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: generate
Source: https://github.com/gonzaloetjo/setup-claude-md/tree/main/skills/generate
Command: npx skills add https://github.com/gonzaloetjo/setup-claude-md --skill generate-gonzaloetjo

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

This Skill improves an existing CLAUDE.md by stress-testing it with fresh agents that plan realistic development tasks and reveal missing project context, so the documentation becomes more reliable for future coding work.

Core Features & Use Cases

  • Iterative refinement: Runs multiple passes, each with a different task archetype, to uncover gaps in commands, conventions, testing, and architecture.
  • Resume-aware workflow: Supports saved state so interrupted generation runs can continue from the last completed iteration.
  • Quality pruning: Keeps the CLAUDE.md focused by removing anything Claude can infer from the codebase and preserving only actionable, high-value guidance.
  • Use case: A team has a basic CLAUDE.md from /init, but complex repo-specific workflows keep getting missed; this Skill deepens the file through realistic planning exercises.

Quick Start

Use the generate skill to refine the repository’s existing CLAUDE.md with five iterations of stress-test refinement.

Frequently Asked Questions about generate

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

FAQPage Schema
How do I refine CLAUDE.md to include missing repository conventions and commands?

You can refine CLAUDE.md by iteratively stress-testing it with fresh agents that plan realistic development tasks, revealing missing project context for commands, conventions, architecture, and testing workflows.

What is the best way to improve project documentation generated by /init for complex workflows?

Improving project documentation involves running multiple stress-test passes with different task archetypes to uncover gaps in repo-specific workflows, then applying pruning rules to keep only actionable, high-value guidance.

Does CLAUDE.md refinement work if my repository only has a basic initial file?

Yes, CLAUDE.md refinement works on repositories that already have a base file from /init and need stronger documentation through iterative stress-testing with fresh agents.

Can I resume a CLAUDE.md generation workflow if the process gets interrupted?

Yes, the generation workflow supports resume-aware state handling, allowing interrupted refinement runs to continue from the last completed iteration without losing progress.

How does stress-testing keep CLAUDE.md focused and actionable?

Stress-testing keeps CLAUDE.md focused by applying quality pruning rules that remove anything Claude can infer from the codebase, preserving only actionable, high-value instructions.

When do I need to stress-test CLAUDE.md with multiple task archetypes?

You need to stress-test CLAUDE.md with multiple task archetypes when complex repo-specific workflows keep getting missed and your existing documentation requires deeper, realistic planning exercises.