Engineering and security notes on task-scoped authorization for AI agents.https://agentsfence-alpha.vercel.app/https://agentsfence-alpha.vercel.app/blog/introducing-agentsfence/https://agentsfence-alpha.vercel.app/blog/introducing-agentsfence/How the friend-test alpha of this Hermes plugin turns user intent into a task-scoped authorization policy.Wed, 23 Sep 2026 16:00:00 GMThttps://agentsfence-alpha.vercel.app/blog/the-plan-is-not-the-policy/https://agentsfence-alpha.vercel.app/blog/the-plan-is-not-the-policy/Why AgentsFence gets authority from the user's intent instead of the agent's plan, and why that makes agents both safer and more capable.Wed, 23 Sep 2026 15:00:00 GMThttps://agentsfence-alpha.vercel.app/blog/anatomy-of-a-blocked-prompt-injection/https://agentsfence-alpha.vercel.app/blog/anatomy-of-a-blocked-prompt-injection/Following one injected email through AgentsFence: what the compiler saw, what hardening removed, which rule fired, and what ended up in the audit log.Wed, 23 Sep 2026 14:00:00 GMThttps://agentsfence-alpha.vercel.app/blog/your-guardrail-hook-probably-fails-open/https://agentsfence-alpha.vercel.app/blog/your-guardrail-hook-probably-fails-open/What we found reading Hermes' plugin hook contracts line by line, and seven places a naive guardrail would silently let tools run — including one a security review caught.Wed, 23 Sep 2026 13:00:00 GMT