neat-freak

Reconcile agent memory and project documentation with the current codebase state.

Updated May 26, 2026
One-click install
npx skills add https://github.com/Peiwvy/myskills --skill neat-freak-peiwvy
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: neat-freak
Source: https://github.com/Peiwvy/myskills/tree/main/neat-freak
Command: npx skills add https://github.com/Peiwvy/myskills --skill neat-freak-peiwvy

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

Prevents AI and human teammates from operating on stale, conflicting, or bloated project knowledge by reconciling agent memory and project documentation with the current codebase state.

Core Features & Use Cases

  • Knowledge reconciliation & anti-rot editing: Reviews and edits CLAUDE.md/AGENTS.md, README, and docs/ to merge duplicates, fix outdated rules, and remove obsolete history.
  • Audience-correct documentation: Ensures “rules for the next agent” live in CLAUDE.md/AGENTS.md while “how outsiders/users should use the project” stays in docs/ and README.
  • Anti-overgrowth guardrails: Enforces size limits and avoids stuffing historical narratives into CLAUDE.md, keeping prompts usable and actionable.
  • Cross-layer consistency checks: Handles scope mapping via a change-impact matrix and highlights cross-project doc alignment requirements.

Quick Start

Ask your AI to run neat-freak by saying: sync up the project knowledge across memory, CLAUDE.md/AGENTS.md, README.md, and docs/ by reconciling anything stale or contradictory with the current codebase.

Frequently Asked Questions about neat-freak

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

FAQPage Schema
How do I keep project documentation and agent memory synced with the codebase?

To keep project documentation and agent memory synced, reconcile CLAUDE.md, AGENTS.md, README, and docs against the current codebase state to merge duplicates, fix outdated rules, and remove obsolete history.

What causes knowledge rot in AI agents and how is it prevented?

Knowledge rot happens when agent memory and project documentation become stale or conflicting. It is prevented by applying an impact-mapping approach at end-of-session handoffs to edit or delete obsolete information across memory and docs.

How do I clean up CLAUDE.md and AGENTS.md without bloating the files?

Clean up CLAUDE.md and AGENTS.md by applying anti-expansion rules that enforce size limits and keep boundary rules actionable, preventing historical narratives from stuffing prompts and causing overgrowth.

When should I run a documentation consistency check across my project?

Run a documentation consistency check during end-of-session handoffs or when users report stale docs and conflicting memories, using a change-impact matrix to highlight cross-project doc alignment requirements.

Do I need any dependencies to perform knowledge cleanup across docs and memory?

No dependencies are required to perform knowledge cleanup. You can run the reconciliation process directly to enumerate relevant files and decide what to edit or delete based on the current codebase state.

What is the best way to separate agent rules from user documentation?

The best way to separate agent rules from user documentation is audience-correct editing: keep rules for the next agent in CLAUDE.md or AGENTS.md, and place how outsiders should use the project in docs and README.