GH Repository · uber
ADR
ADR secures enterprise AI agents through observability, security benchmarking, and threat detection. Deployed at Uber.
- stars
- 1,570
- 30-day movement
- +176/day
- Related entries
- 60
- Connections
- 1
Pythonagent-securityai-agentsai-securitybenchmarkagent-appllm-securitymcpmodel-context-protocolprompt-injectioncodexthreat-detectioncursorclaudeclaude-code
ADR is a Python tool from Uber for securing enterprise AI agents through observability, security benchmarking, and threat detection. It is Apache-2.0 licensed and topics indicate coverage of prompt-injection detection and MCP-related security.
Use it when you need observability, benchmarking, and threat detection for AI agents in one place.
Use it to
- Observe enterprise AI agent behavior
- Benchmark agent security posture
- Detect threats like prompt injection
- Assess security of MCP-based agent setups
For Teams deploying AI agents in enterprise settings
- Role
- agent-app
- Language
- Python
- Licence
- Apache-2.0
- Forks
- 148
- Open issues
- 2
- Last push
- 2026-09-16
- Latest release
- sensor-v1.0.0 · 2026-07-31
topicsagent-securityllm-securitythreat-detectionbenchmarkprompt-injectionmcp