Aegis
Six-gate governance for AI agents: PROCEED/PAUSE/HALT decisions with hash-chained audit trails.
- Transport
- Not stated
- Package
- —
- Registry id
- com.undercurrentholdings/aegis
No install snippet on purpose. A working MCP config is a command, its arguments and an environment block — the last two are where API keys live, so this catalogue never stores them and cannot publish them. Follow the link above for the authors' own instructions.
AEGIS evaluates engineering proposals through six gate checks over caller-supplied metrics and records each authenticated decision in a hash-chained audit trail. Every proposal is judged against six named gates (risk, profit, novelty, complexity, quality, utility) and yields a structured decision of PROCEED, PAUSE, HALT, or ESCALATE with confidence scores, rationale, and next steps. The product can be used through several integration options: a Python SDK client, a CLI, an MCP server, and a GitHub Action.
AEGIS uses hybrid post-quantum cryptography: classical Ed25519 combined with ML-DSA-65 (FIPS 204) for signatures, and ML-KEM-768 (FIPS 203) for encryption. It supports shadow mode and drift detection, evaluating without enforcing decisions, using KL-divergence to compare decision patterns against a baseline.
Summary drafted from the project's own website. Every sentence is backed by text on that page and was reviewed before publishing.