What is the 3-Tier documentation system?

It is a structured approach to project knowledge where Tier 1 covers high-level architecture, Tier 2 covers specific components, and Tier 3 covers granular feature details.

How does the cross-reference validation work?

It scans markdown files for internal links using specific patterns (like @path/to/file) and verifies that the referenced files actually exist in the file system.

Why is there a 200-line limit for Tier 1 files?

Keeping core files like CLAUDE.md under 200 lines ensures the AI can quickly ingest high-level context without wasting tokens or getting lost in excessive detail.

Does this skill check every markdown file in the repository?

The skill utilizes smart change detection via git to prioritize modified files, though it can perform a full verification if changes cannot be detected.

Can this skill identify circular references?

Yes, it includes specialized checks to detect self-references and circular paths that could confuse AI agents during context retrieval.

3-Tier Documentation Verifier

Name: 3-Tier Documentation Verifier
Author: changoo89

bychangoo89

•

Content Management

Validates project documentation against 3-Tier system standards to ensure context efficiency and structural integrity.

The Documentation Verification skill automates the maintenance of a high-performance 3-Tier documentation system within Claude Code environments. It enforces strict structural rules, such as line limits (≤200 lines for Tier 1 files) and specific file counts, while identifying broken cross-references and circular dependencies. By ensuring documentation remains concise and well-linked, it prevents context bloat and helps AI assistants navigate project knowledge more effectively without exceeding token limits.

Key Features

01Circular reference identification to prevent documentation loops

02Validation of mandatory 3-Tier file structure and counts

03Real-time detection and reporting of broken cross-references

044 GitHub stars

05Automated Tier 1 line limit enforcement (≤200 lines per file)

06Smart change detection to run verification only on modified markdown files

Use Cases

01Pre-commit checks to prevent merging oversized or broken documentation files

02Standardizing architecture files like CLAUDE.md and project-structure.md for AI ingestion

03Post-sprint cleanup to ensure documentation remains within AI context limits

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add changoo89/claude-pilot docs-verify

For use in Claude.ai and ChatGPT

Download Skill