ToolGuardian secures agent-tool interactions through pre-admission vetting and task-aware runtime authorization. It progressively converts tool descriptions, system-call traces, mock-execution effects, and source analysis into structured facts, then applies Answer Set Programming policies over capabilities, effects, task context, conformance, and multi-tool composition. On 16 MCP-style tools, including eight malicious variants, ASP achieved 0.86 deny-class F1 and 88% vetting accuracy using description, syscall, and observed-effect evidence. Fully specified runtime realizations classified all 20 reported scenarios correctly, while rule ablations reduced performance.
No heat snapshots are available in the last 24 hours.