https://amitnaik.com/tags/agent-security/ Recent content in Agent Security on Amit Naik Hugo -- 0.144.2 en-us Mon, 28 Sep 2026 08:59:55 -0700 https://amitnaik.com/til/2026-09-28-openhunterai-an-open-red-team-engine/ Mon, 28 Sep 2026 08:59:55 -0700 https://amitnaik.com/til/2026-09-28-openhunterai-an-open-red-team-engine/ This post introduces OpenHunterAI, a security red-team engine released as open source after its creators abandoned their startup plan. The repository is a starting point for examining how AI-assisted security testing is implemented. https://amitnaik.com/til/2026-09-28-ai-routers-introduce-another-trust-boundary/ Mon, 28 Sep 2026 08:59:45 -0700 https://amitnaik.com/til/2026-09-28-ai-routers-introduce-another-trust-boundary/ This security research examines risks introduced by intermediary AI routers. The authors report attacks and failures in the routing ecosystem. The architectural takeaway is to include intermediaries in the threat model instead of assuming a familiar model name guarantees an unchanged execution path. https://amitnaik.com/til/2026-09-28-mapping-attack-paths-across-ai-tools/ Mon, 28 Sep 2026 08:59:34 -0700 https://amitnaik.com/til/2026-09-28-mapping-attack-paths-across-ai-tools/ AgentHound models relationships across AI and MCP systems to expose potential attack paths. The bookmark is useful for thinking about security as a connected system of identities, tools, and permissions rather than a checklist for one model endpoint. https://amitnaik.com/til/2026-09-28-a-reported-attack-through-poisoned-skill-instructions/ Mon, 28 Sep 2026 08:59:29 -0700 https://amitnaik.com/til/2026-09-28-a-reported-attack-through-poisoned-skill-instructions/ The author reports an incident in which malicious instructions in a SKILL.md file caused downloaded malware to reappear. It is a personal incident account, not an independently verified investigation. The lesson is to treat retrieved skill instructions as executable influence over an agent's behavior.