Pinned Loading

1. agentlock agentlock Public

   An adversarially benchmarked reference implementation for pre-action AI agent authorization. Provenance-based gating for LLM agent tool calls: deny-by-default permissions, parameter lineage, signed…

   Python 19 3

2. agentprov agentprov Public

   Vendor-neutral benchmark measuring provenance attribution in LLM agents: does a defense correctly identify where an injected instruction originated as it launders through context? Scores any defens…

   Python

3. awesome-ai-agent-attacks awesome-ai-agent-attacks Public

   A curated timeline of real AI agent security incidents, breaches, and vulnerabilities (2024-2026). Every entry sourced and dated.

   64 6

4. crewai-agentlock crewai-agentlock Public

   Per-tool authorization for CrewAI agents. Every tool call gated, logged, and cryptographically signed.

   Python