gemini-agent-sdk

Pass

Audited by Gen Agent Trust Hub on Aug 17, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill instructions are purely architectural and focus on maintaining parity between different LLM runner implementations. No malicious patterns were identified in the instructions or evaluation cases.- [COMMAND_EXECUTION]: No executable commands or shell scripts are present. References to 'Gemini CLI' and 'runner.ts' serve as architectural guidance for project organization rather than immediate execution commands.- [PROMPT_INJECTION]: Formatting constraints, such as omitting preambles and limiting bullet points, are standard output control instructions and do not attempt to bypass safety filters or override core system instructions.- [DATA_EXFILTRATION]: No network access, sensitive file path access, or credential exposure patterns were identified. The referenced file paths are local to the hypothetical project structure.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 17, 2026, 11:52 AM
Security Audit — agent-trust-hub — gemini-agent-sdk