skills/sanjay3290/ai-skills/imagen/Gen Agent Trust Hub

imagen

Pass

Audited by Gen Agent Trust Hub on Sep 15, 2026

Risk Level: SAFEDATA_EXFILTRATIONPERSISTENCEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [DATA_EXFILTRATION]: The skill communicates with the Google Gemini API (generativelanguage.googleapis.com) to transmit user prompts and receive image data. This is standard functionality for an image generation tool and targets a well-known service domain.
  • [PERSISTENCE]: Documentation in README.md and reference.md provides manual instructions for users to persist their GEMINI_API_KEY by appending export statements to shell profile files such as ~/.zshrc or ~/.bashrc. This is a standard configuration practice and requires manual user action.
  • [INDIRECT_PROMPT_INJECTION]: The skill represents a surface for indirect prompt injection as it processes user-provided prompts and has the capability to write files to the local system.
  • Ingestion points: User prompts provided as command-line arguments to scripts/generate_image.py.
  • Boundary markers: None; the prompt is directly inserted into the JSON request body sent to the Gemini API.
  • Capability inventory: The script can write binary data (PNG images) to the local filesystem using Path.write_bytes.
  • Sanitization: The input is JSON-encoded before transmission, but there is no specific filtering of the prompt content itself, relying on the provider's safety filters.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 15, 2026, 11:16 AM
Security Audit — agent-trust-hub — imagen