Reference execution layer for governed AI agent actions, with bounded synthetic destinations, runtime adapters, and retained execution evidence.
Evidence registry for scientific claims, evidence commitments, provenance, and revision history. Includes a local blockchain reference implementation.
A minimal runtime control and replay layer for AI agents. It records proposed actions, authorization decisions, blocked actions, external-tool reliance, and replayable run traces before agents affect enterprise systems.
Binary attestation protocol and reference verifier for Ed25519 signatures, issuer provenance, and explicit trust snapshots.
Open, vendor-neutral record format for portable decision evidence in AI-mediated and cross-boundary decisions
A public schema and reference validator for declaring what actions an AI agent may propose, what tools it may call, and what authority each action requires.
Independent verifier for Cognous Open Control Stack decision records. It rechecks a run from retained records using its own parser and canonicalization, with no imports from the stack, and reports what it could and could not confirm. Report-only. Not a certification.
Engineering integration and evidence coordination for Cognous Open Control Stack, with pinned component interfaces and bounded synthetic qualification tests.
A public JSON schema, validator, and example format for packaging AI-agent runs into portable replay bundles for audit, review, redaction, signing, and incident reconstruction.
Open reference infrastructure for AI agent governance: runtime authorization, constrained execution, provenance, replay, and decision evidence.
Ergonomics package for Cognous Open Control Stack.
Experimental governed messaging reference connecting recipient assessment, action authorization, bounded execution, and retained decision evidence.
Alvorada: an open institutional governance framework with constitutional designs, proportional oversight, authority contracts, and domain guidance.
A public template and schema set for converting AI-agent runtime records into business-facing governance evidence for security, risk, compliance, and audit review.
Reproducible cross-component tests and qualification evidence for Cognous Open Control Stack, with pinned component versions, automated runs, results, and documented limitations.
To associate your repository with the cognous-ocs topic, visit your repo's landing page and select "manage topics."