agents-secure

This skill audits deployed NeMo agents for security vulnerabilities by inspecting configurations for missing guardrails and sampling telemetry data to detect PII exposure and leaked credentials, providing structured remediation suggestions.

Detecting leaked secrets in AI agent telemetry requires regex-based analysis to scan for sensitive information like API keys and SSNs. This skill samples agent telemetry data to uncover leaked credentials and sensitive information automatically.

Checking if content-safety guardrails are active on model endpoints requires automatically inspecting agent configurations. This skill performs security auditing to ensure content-safety guardrails are active on model endpoints and flags missing protections.

Yes, you can scan agent configurations for security vulnerabilities without manual inspection. This skill automatically inspects deployed NeMo agents to identify missing guardrails, PII exposure, and leaked secrets in telemetry data.

To get remediation suggestions for rotating credentials and anonymizing sensitive data, run this security audit skill. It provides clear, step-by-step suggestions for applying guardrails, rotating credentials, and anonymizing sensitive data.

Yes, this security audit persists snapshots for ongoing compliance monitoring of NeMo agents. It provides persistence of security snapshots alongside structured remediation suggestions to maintain long-term safety posture.