gemini

Pass

Audited by Gen Agent Trust Hub on Mar 9, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection attacks because it processes user-provided input strings directly through the Gemini CLI.\n
  • Ingestion points: Data enters the agent context through positional prompt arguments passed to the gemini command as seen in SKILL.md.\n
  • Boundary markers: There are no explicit delimiters or boundary markers specified in the documentation to isolate external content from system instructions.\n
  • Capability inventory: The skill utilizes the gemini command-line utility for text generation and analysis.\n
  • Sanitization: No sanitization, filtering, or validation of user-provided prompts is implemented or suggested.\n- [COMMAND_EXECUTION]: The skill executes the gemini binary on the system to perform its primary functions. This is the intended behavior for the skill's operational purpose.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 9, 2026, 09:12 PM