flow-doc-sync

Compare code behavior against documentation and generate a severity-tagged drift report.

25|3|Updated Feb 19, 2026
One-click install
npx skills add https://github.com/benkruger/flow --skill flow-doc-sync
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: flow-doc-sync
Source: https://github.com/benkruger/flow/tree/main/skills/flow-doc-sync
Command: npx skills add https://github.com/benkruger/flow --skill flow-doc-sync

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Full codebase documentation accuracy review. Compares behavioral sources (skills, lib scripts, config files) against all documentation surfaces (README.md, docs/**/.md, CLAUDE.md, .claude/rules/.md) and produces a severity-tagged drift report. Read-only — reports drift but does not fix anything.

Core Features & Use Cases

  • Drift detection across code and documentation surfaces to surface mismatches.
  • Severity-tagged drift report with actionable findings.
  • Read-only auditing that helps maintain documentation accuracy over time.

Quick Start

Run the flow-doc-sync skill to generate a drift report comparing code behavior with documentation.

Frequently Asked Questions about flow-doc-sync

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

FAQPage Schema
How do I detect documentation drift between my code and README files?

Documentation drift is detected by comparing behavioral sources like scripts and config files against all documentation surfaces, generating a severity-tagged report that identifies mismatches without modifying code.

What is the best way to check for stale documentation across a full codebase?

Checking for stale documentation involves scanning all markdown surfaces including README.md and docs directories, then comparing them against actual code behavior to produce an actionable drift report with severity tags.

Can I audit code and documentation mismatches without modifying my files?

Yes, you can audit documentation accuracy in read-only mode, which reports drift findings like STALE, MISSING, and OUTDATED entries without altering any code or documentation files.

How do I generate a drift report for CLAUDE.md and project rules?

To generate a drift report, the tool scans behavioral sources and compares them against documentation surfaces including CLAUDE.md and .claude/rules directories, outputting inline findings and a completion banner.

What types of findings are included in a documentation drift report?

A documentation drift report includes severity-tagged findings categorized as STALE, MISSING, or OUTDATED, highlighting specific mismatches between documented behavior and actual code implementation.

Does documentation drift detection work with markdown files in nested docs directories?

Yes, documentation drift detection scans all markdown surfaces including nested paths like docs/**/*.md, ensuring comprehensive coverage across the entire codebase documentation structure.