What happened
On Oct 15 BlueRock (agentic operations/Security) launched the Agentic Trust Registry: automated analysis of AI agent, skill, and MCP server code producing 0-100 trust scores across 34 rules (MCP tool permissions, prompt/metadata injection, OAuth/PKCE correctness, skill provenance/rug-pull, secrets, SSRF, over-privileged tools), claiming 2-10x faster than existing scanning tools.
Why it matters
As enterprises pull third-party agents, skills, and MCP servers off marketplaces, they lack a vetting signal for malicious/over-privileged code. A public trust registry with transparent scoring rules is a new, needed supply-chain trust signal for the agent ecosystem, complementing enterprise vetting tools like Tenable Exchange Inspector.
Applicability
Platform teams and security teams adopting third-party agents/skills/MCP servers should use the registry's trust scores and detailed rule reports as a pre-adoption screening layer.