paw-cra-design-batch
Pass
Audited by Gen Agent Trust Hub on Apr 1, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill fetches generated visual assets from the fal.ai API services.
- [COMMAND_EXECUTION]: The production loop executes shell-based curl commands to interact with the fal.ai generation queue and potentially uses Puppeteer/Playwright for template rendering.
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests untrusted content from campaign briefs and calendars which is then interpolated into prompts for the AI image generation model.
- Ingestion points: Parsed campaign briefs and content calendars defined in
references/brief-intake.md. - Boundary markers: Not explicitly defined in the intake or generation protocols.
- Capability inventory: Local file system writes for assets and logs, network requests via
curltofal.run, and automated rendering via browser-based tools. - Sanitization: No specific sanitization or filtering logic is described for the incoming brief data.
Audit Metadata