exact-text-image

Pass

Audited by Gen Agent Trust Hub on Aug 12, 2026

Risk Level: SAFE
Full Analysis
  • [DATA_EXPOSURE_AND_EXFILTRATION]: The script stores the AI Hive API key in ~/.ai-hive/config.json. It proactively secures this file by setting permissions to 0600, ensuring that only the current user has read and write access to the credentials.
  • [INDIRECT_PROMPT_INJECTION]: The skill accepts user-defined prompts to control image generation.
  • Ingestion points: Untrusted data enters via the --prompt argument in scripts/imagegen.py.
  • Boundary markers: None; the prompt is encapsulated in a JSON request body.
  • Capability inventory: Performs network requests to ai-hive.iclip.cn and writes generated image files to the local ~/Downloads/AiHive directory.
  • Sanitization: No client-side sanitization is performed, relying on the service's API-level safety filters.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 12, 2026, 12:19 PM
Security Audit — agent-trust-hub — exact-text-image