Request a signed review.

Tell us about the agent. We reply with a fixed price for that agent and a review window. Nothing is accessed until you approve the scope.

1. Scope. You describe the agent, its platform, tools and data. We reply with a fixed price and the access needed, read-only by default.
2. Review. We inventory the agent's identity, permissions, MCP tools and data reach, then attack its real configuration: poisoned tools, indirect prompt injection, over-broad permissions, exfiltration.
3. Proof. You get every finding with its proof level and verdict, the configuration fixes, an evidence pack, alerting rules for your SIEM and a signed go/no-go.

That did not go through. Check the required fields and try again.