gemini-3-pro-image-generator

Pass

Audited by Gen Agent Trust Hub on Mar 25, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes a Bash script provided in SKILL.md that uses curl, jq, and base64 to communicate with image generation services and process binary data.\n- [EXTERNAL_DOWNLOADS]: Communicates with Google's Generative Language API (generativelanguage.googleapis.com) to download generated image content, which is a well-known and expected service for this task.\n- [PROMPT_INJECTION]: The skill exhibits a surface for indirect prompt injection as it interpolates user-provided text directly into a JSON payload for a shell command.\n
  • Ingestion points: User-provided prompt variable in SKILL.md.\n
  • Boundary markers: None; the prompt is placed directly within the JSON string.\n
  • Capability inventory: Network access (curl) and local file system writes (>).\n
  • Sanitization: None; the prompt is not escaped or validated, allowing a crafted prompt to potentially alter the API request schema or parameters.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 25, 2026, 12:14 AM