Agent Identity

Verification notes: New listing, added this pass. Announced August 4, 2026 at the Black Hat Conference — confirmed via Rubrik's own newsroom press release (Rubrik trades on NYSE as RBRK). Agent Identity is an expansion of Rubrik Agent Cloud, joining three existing pillars (Agent Observability, Agent Runtime Security, Agent Rewind) and works by minting scoped, short-lived tokens per individual tool call rather than issuing standing credentials — every MCP tool call clears a behavioral-analysis check (via Rubrik's SAGE governance engine), an access-policy check, and an identity-verification check before execution. mcp is flagged true based on Rubrik's own description of monitoring 'every agent and model context protocol (MCP) at runtime.' No A2A support is mentioned in the release, so that flag stays false rather than assumed. Rubrik's own release states Agent Identity 'integrates seamlessly with Okta and Microsoft Entra ID' — see Okta for AI Agents and Entra Agent ID — extending those existing identity systems to cover agent actions rather than replacing them. cryptoSigned stays false: the release describes short-lived scoped tokens but doesn't specifically characterize them as cryptographically signed the way Entra Agent ID or Skyfire are documented. No independent third-party audit of Agent Identity itself has been located. See also AIR, a newer listing with an overlapping but distinct focus — both grant or vet access at the individual tool-call level rather than issuing standing credentials, though Agent Identity governs the enterprise's own agents while AIR is oriented more toward vetting third-party skills, MCP servers, and plugins before they're installed. Re-checked this pass: no material change to the flags found; product remains branded 'Agent Identity' with the architecture described above unchanged per Rubrik's own blog.