Topic: agent-security-eval

4 repositories

scadastrangelove/agent-audit

Forensic auditor for local AI coding agents (Claude Code, Codex CLI, OpenClaw) and project-surface scanner for repos containing skills, plugins, and MCP manifests. Reads session logs, configs, and instruction files, detects known-bad patterns using 296 bundled rules in total.

★ 15PythonForks 3

HPBhoir/AgentSecBench

Open-source framework for evaluating the security of tool-using AI agents.

★ 0PythonForks 0

psandhir/horustrace

Open-source static security analysis for agentic systems, with dependency graphs, capability tracing, MCP analysis and attack-path detection.

★ 0PythonForks 0