Repo

CUSTOS-CORE

@ceyptoslimby Tavares Heath · Independent

0

Deterministic, fail-closed execution gate for AI agents. Every proposed action is evaluated against explicit per-tenant policy before it runs (deny by default, tool allowlists per tenant). Every decision, allow or deny, goes into a hash-chained, tamper-evident audit record — with chain heads anchored in external custody (a git repo plus RFC 3161 timestamps), so a consistent rewrite of the store is caught by the anchors, not just the chain's own verifier. The gate is not a model, so an agent can't reason or prompt its way past it. AGPL open core. Proof: 550 CI tests, plus red-proof CI — four sabotage classes and three custody attacks on a weekly schedule, where a deliberate weakening of any boundary must turn the suite red; run records are frozen in the repo. An independent researcher reconstructed the evaluation records, re-hashed every artifact, and published his verification: [github.com/factnotebook/public-reconstructions](https://github.com/factnotebook/public-reconstructions)

Built withPythonFastAPIPostgreSQLKubernetesHelmGitHub ActionsAI AgentsSecurityGovernanceOpen Source

See what other members are building and what each project is looking for.