gcloud-log
Pass
Audited by Gen Agent Trust Hub on May 7, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill relies on executing gcloud CLI commands to interact with Google Cloud services, including reading log data, listing projects, and switching accounts. These commands are defined in the SKILL.md instructions and implemented using the subprocess module in the refresh-context.py script.
- [PROMPT_INJECTION]: The skill presents an indirect prompt injection surface (Category 8) due to its processing of external, untrusted log data.
- Ingestion points: Log payloads (textPayload and jsonPayload.message) are retrieved from GCP and processed by the agent.
- Boundary markers: The skill does not define explicit delimiters or instructions to prevent the agent from executing commands embedded within the fetched logs.
- Capability inventory: The agent has access to shell execution via the Bash tool (SKILL.md) and Python's subprocess.run (refresh-context.py).
- Sanitization: Log content is not sanitized or validated before being analyzed by the agent, although structured formatting is applied via CLI flags.
Audit Metadata