How does this skill verify that documentation is accurate?

The skill uses Read, Grep, and Glob tools to search the actual codebase for evidence. For example, it checks if documented API routes exist in controllers or if architectural layers match the directory structure.

What is the difference between scope alignment and fact accuracy?

Scope alignment checks if the document is talking about what it's supposed to (relevance), while fact accuracy verifies if what is written matches the current state of the code (truth).

Which specific documents does the auditor target?

It targets core project files such as CLAUDE.md, README.md, architecture.md, tech_stack.md, api_spec.md, and runbooks, while excluding task logs and reference folders.

Can I use this skill as part of an automated pipeline?

Yes, it is designed as a worker skill (L3) intended to be called by a coordinator like ln-600 to provide structured JSON feedback for a full project audit.

Semantic Documentation Auditor

Name: Semantic Documentation Auditor
Author: levnikolaevich

bylevnikolaevich

•

113

•

Learning & Documentation

Verifies technical documentation accuracy by cross-referencing document claims against the live codebase state.

The ln-601-semantic-content-auditor is a specialized L3 worker skill designed to maintain high-quality, reliable project documentation. It performs deep semantic audits by extracting a document's intended scope and validating its content against the actual codebase using tools like Grep and Glob. It identifies outdated paths, incorrect API specifications, architectural mismatches, and scope creep, providing a detailed scoring report with actionable fix suggestions. This skill ensures that your documentation remains a 'source of truth' rather than a collection of stale information.

Key Features

01Structured reporting with severity levels, evidence, and line-specific fixes

02113 GitHub stars

03Scope alignment analysis to detect off-topic content or missing coverage

04Automated fact-checking of documentation against source code and file structures

05Quantitative scoring for both scope alignment and factual accuracy

06Validation of API endpoints, tech stack versions, and architectural patterns

Use Cases

01Validating architecture.md and API specs against current implementation logic

02Ensuring runbooks and CLAUDE.md contain valid file paths and working commands

03Auditing project documentation during handovers or major version updates

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add levnikolaevich/claude-code-skills ln-601-semantic-content-auditor

For use in Claude.ai and ChatGPT

Download Skill