documentation-drift

Portable, always-fresh docs for agents and humans. A Surface hub is a conformant OKF concept; Surface adds the freshness OKF leaves out - it fails the build when the code a doc describes changes. Deterministic, no LLM.

A context-aware AI code review bot that acts like a Senior Engineer. Features 2-stage triage, documentation drift detection, and architectural enforcement.

Generate docs from your code graph and docstrings, then gate them in CI so they can never go stale.

Finds documentation that no longer matches your code. Deterministic AST check for pull requests; no API key. TypeScript, JavaScript, Python.

🔍 AI-powered Code↔Docs Drift Detector and Security-first Auditing Pipeline built for the Anna Hackathon

A pi prompt template that keeps CLAUDE.md in sync with your codebase — generates it on first run, then audits drift and updates it surgically as your project evolves

Tracks documentation freshness vs code churn to identify outdated AI agent context

Keeps architecture diagrams honest — binds D2 diagram nodes to real repo paths and fails CI when they diverge

Evidence-backed documentation drift control for agentic repositories.

GitHub Action that automatically flags stale documentation by detecting drift between code changes and project docs using LLMs.

Rust CLI that finds drift between Claude skill prose and installed tool surfaces, and parks review-gated fix proposals.

Reads the promises a project makes to its customers and checks whether anything in the code keeps them.

Finds documentation that has quietly stopped being true - and reports what it could not determine. Drift detection + git-versioned architecture memory (OKF) for AI coding agents.

Detect verifiable documentation drift in TypeScript repositories.

Generate a CLAUDE.md where every line names the file it came from. No model, no network, no tokens.

Claude Code skill for architecture reviews, docs-to-code drift detection, and evidence-based technical debt analysis across 12 dimensions — scalability, extensibility, reliability, security, change safety, data architecture, observability, and more.

Blocks a push when your documentation can be proven wrong. Reads your .gitignore, learns your repo layout, reports evidence instead of staleness scores. Python 3.8+, stdlib only, zero config.