domain

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (medium risk: 0.30). The workflow ingests outsider-authored free text at runtime via user-provided candidate/word inputs and then reads public-domain text from arbitrary third-party sites by fetching taken domains’ HTML/lander bodies in Step 7a (using curl ... https://$domain/ and grepping markers).

MEDIUM W009: Direct money access capability detected (payment gateways, crypto, banking).

• Direct money access detected (high risk: 1.00). The skill includes explicit commands/APIs to purchase domains (directly charge money): "vercel domains buy .com" and a Namecheap API domains.create curl call, and it references NAMECHEAP API credentials. Even though it prompts to confirm with the user, these are concrete, actionable financial operations (execute a paid purchase), so the skill grants direct financial execution capability.