agent-security-monitor

To scan for exposed secrets and insecure SSH keys, run the security monitoring script against your agent workspace. It checks environment files, skill manifests, and SSH key permissions, then outputs color-coded alerts and detailed logs for quick triage and audit.

Supply-chain security monitoring for AI agents detects skills lacking proper SKILL.md documentation and flags suspicious patterns. It validates permission manifests to reduce false positives and ensures safe skill ecosystem governance during development and operation.

Yes, you can monitor suspicious commands in agent history using Bash with jq, grep, find, and stat. The monitoring script scans recent command histories and guards log files against leaks to prevent compromised agent workflows.

Yes, this security monitoring script works with standard Bash and requires only jq, grep, find, and stat as dependencies. It implements configurable checks for environment files, skill manifests, and command histories without requiring external security platforms.

Reduce false positives when scanning for unsafe data flows by implementing a permission manifest validation workflow. The monitoring script uses configurable checks to ensure proper masking of sensitive patterns and verify safe skill configurations.