Agentic Identity IndexKeycard agent identity review
Delegated tool access with step-up approval
Agent discovery 3.0 / Per-action policy 9.0 / Agent credentials 9.0
General purpose score 8.10 joint
Where it falls short
- Broad agent discovery is not verified.
- Bring-your-own-cloud deployment is marked coming soon in the deployment documentation.
Documented strengths
- Each agent has an identity evaluated with user, application and workload context.
- Per-request policy evaluates access in runtime context.
- Just-in-time, scoped credentials limit standing secret exposure.
Every score, with its reason
Agent identity and ownership 8.0
Each agent has an identity evaluated with user, application and workload context.
Official source / Verified 2026-11-04
Agent discovery 3.0
Broad discovery of unknown agents is not documented; score reflects registered application identities only.
Official source / Verified 2026-11-04
Per-action policy 9.0
Per-request policy evaluates access in runtime context.
Official source / Verified 2026-11-04
Agent credentials 9.0
Just-in-time, scoped credentials limit standing secret exposure.
Official source / Verified 2026-11-04
Delegation and on-behalf-of access 9.0
User-on-behalf and previously authorised background access are documented.
Official source / Verified 2026-11-04
Audit and lifecycle 8.0
Transaction telemetry and audit export are documented; plan limits apply.
Official source / Verified 2026-11-04
Human approval 8.5
Sensitive actions can require step-up human approval.
Official source / Verified 2026-11-04
MCP gateway and tool control 9.0
The gateway aggregates MCP servers and applies policy to individual tool calls.
Official source / Verified 2026-11-04
Ownership and pricing
Standalone vendor / Ownership source
Starter free; Team $500 per month; Enterprise quote only / Pricing source