agent-guard

Pre-installation scanning detects prompt injection and command injection by validating target skill files against predefined checks. Running this scanner outputs a structured scan result with recommendations, reducing the risk of integrating unsafe external capabilities.

Pre-install skill validation is a safety process that checks third-party plugins for command injection and data exfiltration patterns before integration. It enforces safety by applying predefined checks to ensure safe integration into development pipelines.

Yes, you can use command injection detection in CI/CD pipelines. The scanner applies pre-installation safety checks across environments, validating any plugin or skill integration where third-party capabilities are added during automated development workflows.

You check a plugin for data exfiltration patterns by running a scanner on the target skill file before installation. This detects common command-and-data exfiltration patterns and outputs a structured scan result with actionable safety recommendations.

Yes, pre-install plugin validation works for local development setups. The scanner applies predefined safety checks across any environment where third-party capabilities are added, including both local configurations and CI/CD development pipelines.