security-validator

To validate shell commands for security risks, you can assess proposed actions before execution to receive an ALLOW, REQUIRE_APPROVAL, or BLOCK decision. This evaluation checks for destructive file operations, credential exposure, and unsafe network requests during deployments.

Preventing accidental credential exposure in CI/CD pipelines involves validating proposed actions before they run to detect patterns like .env or SSH access. The process returns structured decisions with risk levels and mitigation recommendations to block risky operations.

Yes, you can automatically block file permission changes and untrusted network requests during automated agent tasks. The validation process evaluates proposed actions and returns a BLOCK outcome with clear reasoning if it detects unsafe permission modifications or requests to untrusted domains.

Security validation for API calls handles risk assessment by analyzing proposed actions and returning an unambiguous decision. It enforces policies by assigning a risk level, providing clear reasoning, and offering actionable mitigation recommendations for each evaluated API call.

You need to require human approval for risky operations in automated deployments when proposed actions present moderate security risks that warrant oversight. The validation mechanism returns a REQUIRE_APPROVAL decision, assigning a risk level and actionable mitigation recommendations before execution proceeds.

Yes, security validation works with deployment scripts that modify file permissions and call external services. It evaluates these proposed actions to determine whether to block, require human approval, or allow automated execution based on detected security risks and untrusted domain requests.