Preview Release: Early-access preview, information and pricing subject to change.

Security & Governance Posture

AGEI supports governance evidence through zero-trust architecture, cryptographic receipts, and fail-closed behavior.

Core Security Principles

Zero Trust Architecture

No implicit trust. Every agent action evaluated against policies. No action executes without explicit authorization.

API Key Security

Agent systems authenticate using organization-scoped API keys. Keys stored encrypted, never logged in plaintext.

Row Level Security

Complete org isolation via Postgres RLS. Organizations can only access their own agents, policies, and governance data.

Archive, Never Delete

Soft delete pattern for all records. Evidence receipts and governance events retained permanently for audit and compliance.

Evidence Receipts

Immutable cryptographic proof of every governance decision. SHA-256 hashes verify integrity and prevent tampering.

Human Oversight

High-risk actions require explicit human approval via HITL queues. Material impact actions require active oversight.

Connector Authentication

Secure Webhook Integration

External agent systems connect to AGEI via secure webhooks with API key authentication:

  • • Organization-Scoped Keys: Each org has unique API keys tied to their database records
  • • Encrypted Storage: API keys encrypted at rest and in transit
  • • Header-Based Auth: X-AGEI-API-Key header required on all connector requests
  • • Request Validation: Webhooks verify connector registration, agent authorization, and policy evaluation
  • • Fail-Closed Behavior: Authentication failures block agent execution immediately

Fail-Closed & Cryptographic Evidence

Fail-Closed Behavior

When AGEI cannot reach a governance decision (missing policy, auth failure, system error), the default is deny. No agent action executes without explicit approval.

SHA-256 Evidence Hashing

Every governance event and evidence receipt includes SHA-256 hash of decision context, policy result, and metadata. Verifiable proof that governance occurred.

Data Protection

What AGEI Stores

  • • Agent Metadata: Name, type, governance classification, context access, authority levels
  • • Policy Definitions: Rules, conditions, decision logic (not customer data)
  • • Governance Events: Timestamps, decisions, policy evaluations, approval flows
  • • Evidence Receipts: SHA-256 hashes, decision outcomes, human review actions
  • • User & Org Info: Account data, org structure, role assignments

AGEI does NOT store: Customer PII, sensitive business data, agent conversation history, or plaintext credentials.

Deploy Secure Agent Governance

Start governing your AI agents with zero-trust security and cryptographic evidence