ZIRAN Agent Security Scan
Version updated for https://github.com/taoq-ai/ziran to version v0.9.0.
- 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 a security testing tool designed to identify vulnerabilities in AI agents, focusing on tool combinations, memory, and multi-step reasoning rather than isolated components. It automates the discovery of attack paths, side-effect detection, and multi-phase campaigns by leveraging graph-based analysis and autonomous pentesting capabilities. ZIRAN addresses complex security challenges such as detecting dangerous tool interactions, cross-agent vulnerabilities, and protocol-specific attacks, making it a comprehensive solution for securing AI systems.
Release notes
What’s Changed
- feat: address three critical benchmark gaps by @leoneperdigao in https://github.com/taoq-ai/ziran/pull/46
- feat: add business impact categorization to findings by @leoneperdigao in https://github.com/taoq-ai/ziran/pull/47
- feat: add AILuminate-style resilience gap metric by @leoneperdigao in https://github.com/taoq-ai/ziran/pull/48
- feat: add utility-under-attack measurement by @leoneperdigao in https://github.com/taoq-ai/ziran/pull/49
Full Changelog: https://github.com/taoq-ai/ziran/compare/v0...v0.9.0