The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The bundled shell script scripts/health-check.sh accepts cluster context and namespace as command-line arguments and interpolates them directly into kubectl command strings without escaping. This creates a potential surface for command injection if malicious inputs are provided. * Ingestion points: Command-line arguments $1 and $2 in scripts/health-check.sh. * Boundary markers: None. * Capability inventory: kubectl command execution. * Sanitization: None present in the script code.

k8s-troubleshoot