Why does the audit recommend keeping CLAUDE.md under 100 lines?

Claude's performance and instruction-following capabilities degrade when the core instruction file is too long. The skill encourages extracting detailed info to the docs/ folder via @import to keep the primary context focused.

Does this skill automatically delete my documentation?

No. It auto-fixes structural order and removes redundant 'fluff' (like 'write clean code'), but any substantial domain-specific content is flagged for developer review or safely moved to modular documentation files.

How does command verification work?

The skill reads your package.json or project config files and cross-references the actual scripts with the instructions in CLAUDE.md to ensure the AI isn't using hallucinated or outdated commands.

What are 'hook candidates' identified by the audit?

These are rules that can be enforced programmatically (like 'always run lint') rather than just being a written instruction. The skill suggests moving these to pre-commit or tool-use hooks for better reliability.

Best Practices Audit

Name: Best Practices Audit
Author: Brite-Nites

byBrite-Nites

0•

内容管理

Audits and auto-fixes CLAUDE.md documentation to ensure project instructions follow Anthropic's official best practices for AI efficiency.

The Best Practices Audit skill maintains the health of your project's CLAUDE.md file by enforcing strict size limits, structural standards, and instruction accuracy. It prevents context drift by validating listed commands against your actual codebase, automatically extracting bloated sections into a modular @import structure, and identifying rules that are better suited as deterministic hooks. By keeping your project documentation high-density and low-noise, this skill ensures that Claude remains maximally effective as your codebase evolves.

主要功能

01Enforces a 100-line limit for optimal AI context processing and performance

02Automatically migrates large sections to external docs using @import syntax

03Validates build, test, and lint commands against package.json scripts

04Identifies deterministic hook candidates to replace manual workflow instructions

05Flags stale file references and missing required project sections for review

060 GitHub stars

使用场景

01Maintaining documentation hygiene during the 'ship' phase of a development cycle

02Optimizing bloated CLAUDE.md files that have become difficult for the AI to parse

03Standardizing documentation structure across multiple repositories in an organization

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add brite-nites/brite-claude-plugins best-practices-audit

For use in Claude.ai and ChatGPT

主要功能

01Enforces a 100-line limit for optimal AI context processing and performance

02Automatically migrates large sections to external docs using @import syntax

03Validates build, test, and lint commands against package.json scripts

04Identifies deterministic hook candidates to replace manual workflow instructions

05Flags stale file references and missing required project sections for review

060 GitHub stars

使用场景

01Maintaining documentation hygiene during the 'ship' phase of a development cycle

02Optimizing bloated CLAUDE.md files that have become difficult for the AI to parse

03Standardizing documentation structure across multiple repositories in an organization

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add brite-nites/brite-claude-plugins best-practices-audit

For use in Claude.ai and ChatGPT