Sentinel
Local repo health monitor that catches docs drifting from code.
18 detectors across 5 languages, 1,300+ tests; 88% precision benchmarked on real-world repos.
Problem
The faster you build with AI, the faster things drift. AI coding tools help you ship code quickly, but they don't notice when your docs, tests, and code start disagreeing with each other. Linters check code, doc tools check formatting, but nothing checks whether your README still describes reality. This compounds as projects grow, quietly confusing every agent and human who touches the codebase.
Approach
Built a modular scanner with 18 detectors covering 5 languages — linting, TODOs, dependency audits, complexity — but the core value is in cross-artifact analysis. Deterministic detectors handle stale references (paths in docs that don't exist) and dependency drift without any model. Deeper semantic checks use a local LLM as a judgment layer for genuinely ambiguous cases. Results are deduplicated across runs and surfaced in a web UI for triage. The entire tool was built using an autonomous AI development workflow across 41 sessions.
Result
Benchmarked at 88% precision and 100% recall on real-world repos with ground-truth evaluation. 18 detectors, 1,300+ tests, web UI, CI pipeline, and multi-language support across Python, JS/TS, Go, and Rust. Open source and available on PyPI. Built across 41 autonomous sessions in 15 days.