Audits that know your code.
Most organisations have more repositories than they have eyes to keep on them. Sentinel is an installable, Nanocoder-driven workflow that runs continuous security and code audits across the repositories you care about, and files what it finds as issues for a human to act on.
You install it into your own organisation, write the rule packs that describe what to look for, and a scheduled GitHub Actions workflow does the pass. Local models are a first-class path, so the audited code never has to leave hardware you own.
Premium capabilities.
Zero paid tiers.
Scheduled Audits
A GitHub Actions workflow sweeps the repositories you point it at on the schedule you set. No hosted service, no GitHub App to authorise.
Rule Packs You Write
Sentinel ships no rules of its own. You describe what matters in the code you actually ship, so the audit is yours rather than a generic checklist.
Findings as Issues
Every finding lands as an issue on the affected repository, written up for a reviewer, with deduplication and suppression built in.
Local Models First
Point Sentinel at Ollama, LM Studio, llama.cpp or MLX on a self-hosted runner and the audited code never leaves hardware you own.
Install in seconds.
Zero complex setup. Available on your favorite package manager.
Why developers choose Sentinel
Your Code Stays Yours
The local-first path is a first-class option, not an afterthought. Nothing has to be shipped to a third-party scanner.
Ecosystem-Specific
A Solana program, a TypeScript API, and a Rust CLI each get the audit they need — because you wrote the pack for each.
Runs In Your Org
One configuration repository inside your organisation. Uninstalling is deleting it; there is no external service to deauthorise.
Read this before you rely on it
- >Sentinel is in alpha and being built toward v1. The CLI,
sentinel.yaml, and the rule pack format can change between releases. - >It is a triage layer, not a substitute for a formal security audit.
- >It is not a SAST replacement — keep running Semgrep or CodeQL alongside it — and it is not a secret scanner.
- >It files issues; it does not open fix-up pull requests.
- >It ships no rule packs of its own. A fresh install audits nothing until you write one.
Supported by
Sponsorship funds the people building open, privacy-respecting AI tools, and keeps every line of it free at the point of use.
Build with us.
Join developers building the future of local-first AI tooling.