cloud-penetration-testing
Fail
Audited by Gen Agent Trust Hub on Feb 27, 2026
Risk Level: HIGHREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADSCREDENTIALS_UNSAFEDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
- [REMOTE_CODE_EXECUTION]: Fetches and runs the Google Cloud SDK installation script using a piped shell command from its official domain.\n- [COMMAND_EXECUTION]: Performs high-privilege operations including system-level tool installation via sudo and remote script execution on cloud virtual machines using native cloud CLI tools.\n- [EXTERNAL_DOWNLOADS]: Retrieves security auditing and penetration testing frameworks from public Python and PowerShell registries, as well as official vendor download sites.\n- [CREDENTIALS_UNSAFE]: Outlines procedures for extracting sensitive authentication tokens from cloud instance metadata services and managed secret stores.\n- [DATA_EXFILTRATION]: Provides instructions for transmitting captured credentials to external infrastructure, such as custom API gateway endpoints used for IP rotation.\n- [COMMAND_EXECUTION]: Implements techniques for maintaining access by creating backdoor service principals and adding users to high-privilege administrative roles.\n- [PROMPT_INJECTION]: The skill processes untrusted input data from files like user lists and IP lists without explicit sanitization, creating an indirect injection surface. Ingestion points: user-provided text files; Boundary markers: None; Capability inventory: Full shell and cloud API access; Sanitization: Not identified in the provided scripts.
Recommendations
- HIGH: Downloads and executes remote code from: https://sdk.cloud.google.com - DO NOT USE without thorough review
Audit Metadata