paw-cra-design-social

Pass

Audited by Gen Agent Trust Hub on Apr 1, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses shell commands including curl and jq to interact with image generation APIs and process JSON results. These operations are restricted to the intended design workflow.
  • [EXTERNAL_DOWNLOADS]: Fetches generated visual assets from the fal.ai platform (fal.run) and downloads them to the local workspace for processing.
  • [PROMPT_INJECTION]: The skill processes user-supplied briefs to generate prompts for AI models. While it lacks explicit sanitization or boundary markers for this input, it is a core function of the design pipeline and presents low risk in this context.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 1, 2026, 02:02 PM