Version updated for https://github.com/yagobski/agentleak to version v0.11.10. This action is used across all versions by ? repositories. Action Type This is a Composite action. Go to the GitHub Marketplace to find the latest changes. Action Summary Summary: AgentLeak is a privacy-leakage testing action designed to audit AI agents’ execution channels, scoring them deterministically, and gating them in CI. It identifies vulnerabilities such as sensitive data leaks through tool calls, shared memory, inter-agent messages, and logs, providing a risk assessment with the AgentRisk metric grounded in GDPR Article 9 and Québec Law 25. The action can be installed via pip and includes local UI and MCP tools for coding agents.

Version updated for https://github.com/yagobski/agentleak to version v0.11.9. This action is used across all versions by ? repositories. Action Type This is a Composite action. Go to the GitHub Marketplace to find the latest changes. Action Summary AgentLeak is a privacy-leakage testing tool designed to audit AI agents across various execution channels. It ensures that sensitive data is not leaked through internal communication and logs by scoring the agent’s performance with the AgentRisk metric, which measures the severity of potential privacy breaches. The Action automates this process within continuous integration pipelines.