@caidongyun

agent-dlp provides data loss prevention across input injection detection, memory integrity, tool approval, output filtering, and audit logging. five functional guards deployed as a pipeline with configurable sensitivity modes.

srs is a task evaluation and role-matching system for security research that gates work admission based on weighted scoring criteria and assigns researchers to tasks via keyword matching.

aagent-system collects ai agent skill samples from multiple sources, performs security scanning for malicious code and credential leaks, and maintains threat intelligence rules. multi-process architecture. trigger: /aagent start.

agent-defender provides static scanning, runtime monitoring, and dlp capabilities for ai agent security. lacks actionable procedures, error handling, and clear intent/outcome definitions.