Popular repositories Loading

1. HYDRACUDA HYDRACUDA Public

   Runtime policy enforcement for AI tool calls. Define a YAML policy, drop it in front of any MCP server or LLM tool layer - every call is allowed, denied, or held for human review before it executes…

   Python 3

2. xtension-registry xtension-registry Public

   Community Registry for Xtension

   1

3. Seideray Seideray Public

   Benchmark for evading tool-call policy enforcement layers in autonomous AI agents

   Python

4. Hekacy Hekacy Public

   Benchmark for credential exfiltration and lateral-use capability in compromised agent environments

   Python

5. Fraegis Fraegis Public

   Benchmark for evaluator-integrity corruption and concealment behavior in AI agent evaluations

   Python