Aegis

Enterprise AI governance workbench — EU AI Act + NIST AI RMF, self-hosted, audit-first.

Aegis catalogs every AI system in your org, classifies each against EU AI Act and NIST AI RMF, evaluates policy-as-code, auto-generates model cards and conformity documentation, and writes every change to an Ed25519-signed append-only ledger.

Financial-services vertical pack included (SR 11-7 model risk, NYDFS Part 500).

Install

pip install aegis-workbench

Or for local development:

pip install -e ".[dev]"

Quickstart

# init an org
aegis init "Acme Bank"

# register an AI system
aegis systems add "Credit Scorer v3" \
    --purpose "consumer credit decisioning" \
    --owner risk-ml \
    --deployment production

# assess it
aegis assess "Credit Scorer v3"

# emit EU AI Act Annex IV technical documentation
aegis report annex-iv "Credit Scorer v3" --out annex_iv.md

# verify the audit ledger
aegis audit verify

Frameworks supported (v1)

• EU AI Act (Regulation (EU) 2024/1689) — risk classification, Annex III use cases, Arts. 9-15 obligations.
• NIST AI RMF 1.0 — Govern, Map, Measure, Manage functions + subcategories.
• SR 11-7 — Federal Reserve model risk management (vertical pack).
• NYDFS Part 500 — cybersecurity obligations for financial services.

Status

Alpha — under active development. See PRD.md for the full scope and roadmap.

License

Apache-2.0