Privatris/AgentLeak
AgentLeak: Open benchmark for privacy leakage in LLM agents — 7 channels, multi-agent, multi-framework.
AgentLeak helps privacy officers and compliance managers identify where confidential information might be exposed when using AI agent systems. It takes various inputs, such as agent outputs, internal messages, or system logs, and reveals if sensitive data like patient records or financial details have been inadvertently leaked. This tool is designed for anyone responsible for data protection and regulatory compliance in organizations deploying AI.
Use this if you need to thoroughly audit your multi-agent AI systems for privacy risks across all internal communication channels, not just the final output.
Not ideal if you are only concerned with privacy leakage in single-agent LLM applications or if you just need to check the final user-facing output.
Stars
9
Forks
1
Language
Python
License
—
Category
Last pushed
Mar 08, 2026
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/agents/Privatris/AgentLeak"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Featured in
Higher-rated alternatives
StonyBrookNLP/appworld
🌍 AppWorld: A Controllable World of Apps and People for Benchmarking Function Calling and...
qualifire-dev/rogue
AI Agent Evaluator & Red Team Platform
microsoft/WindowsAgentArena
Windows Agent Arena (WAA) 🪟 is a scalable OS platform for testing and benchmarking of...
future-agi/ai-evaluation
Evaluation Framework for all your AI related Workflows
agentscope-ai/OpenJudge
OpenJudge: A Unified Framework for Holistic Evaluation and Quality Rewards