kevros - Govern autonomous agent actions
Evaluates agent actions, records provenance, binds intents, verifies outcomes, delegates capabilities, and exports compliance records.
Tags
Updated: 2026-10-07Capabilities
Typical Inputs
Typical Outputs
What this skill does
- Evaluate actions against policies
- Issue cryptographic release tokens
- Attest completed actions
- Bind intents to commands
- Verify intent outcomes
- Export compliance bundles
- Batch governance operations
- Delegate scoped capabilities
- List active delegations
- Revoke active delegations
- Check intent reversibility
Inputs
- Agent ID
- Action type
- Action payload
- Policy context
- Action description
- Intent declaration
- Command payload
- Goal state
- Actual state
- Tolerance
- Time range
- Batch operations
- Delegation scope
- Delegation duration
Outputs
- Policy decision
- Release token
- Verification ID
- Attestation record
- Provenance hashes
- Intent binding record
- Outcome status
- Compliance bundle
- Batch operation results
- Delegation token
- Delegation status
Requirements
- OpenClaw plugin
- Linux, macOS, or Windows
- HTTPS access to governance gateway
- KEVROS_API_KEY or signup access