ZIRAN Agent Security Scan
Version updated for https://github.com/taoq-ai/ziran to version v0.11.0.
- This action is used across all versions by 1 repositories.
Action Type
This is a Composite action.
Go to the GitHub Marketplace to find the latest changes.
Action Summary
ZIRAN is an AI agent security testing tool designed to identify vulnerabilities in AI agents, including those using tools, memory, and multi-step reasoning. It automates the discovery of dangerous tool combinations, detects side-effects from agent actions, and performs autonomous penetration testing with multi-phase campaigns. ZIRAN is framework-agnostic and provides advanced capabilities like graph-based tool chain analysis, multi-agent coordination, and protocol depth testing, addressing security gaps beyond traditional prompt-based tools.
Release notes
What’s Changed
- perf: cache betweenness centrality computation in chain analyzer by @leoneperdigao in https://github.com/taoq-ai/ziran/pull/95
- feat: 100% ALERT benchmark coverage + report failed vector loads by @leoneperdigao in https://github.com/taoq-ai/ziran/pull/128
- feat: expand JailbreakBench coverage to 100% by @leoneperdigao in https://github.com/taoq-ai/ziran/pull/129
- feat: add R-Judge risk scenario coverage (closes #59) by @leoneperdigao in https://github.com/taoq-ai/ziran/pull/130
Full Changelog: https://github.com/taoq-ai/ziran/compare/v0...v0.11.0