Defense-in-depth security toolkit for LLM agents — taint tracking, proxy secret guard, policy engine, and red-team benchmarking

A closed-loop security runtime preventing "The Great Exfiltration" and Indirect Prompt Injection in Autonomous AI Agents.

Signed provenance labels and taint-tracking policy for LLM agent security. The core library behind AgentMesh.

The Security Kernel for AI Agents — MCP/A2A gateway with policy enforcement, taint tracking, sandboxed execution, deterministic envelopes, and Sigstore audit. OWASP ASI 2026 compliant.

Scanner de sécurité PHP par analyse statique : taint tracking, AST tree-sitter, 16 vulns, SARIF/HTML, 89 tests

Identifying Inputs Causing Crash using Input Tagging and Propagation (Taint Analysis)

Make prompt injection structurally unable to cause an unauthorized effect — a Python reference implementation of the Reasoning Kernel (CaMeL-like) pattern.