brightdata

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

• Third-party content exposure detected (high risk: 1.00). The skill's SKILL.md and accompanying scripts (scripts/scrape.sh and scripts/search.sh) explicitly fetch arbitrary public webpages and Google search results via the Bright Data API and return markdown/parsed content that the agent is expected to read and could contain untrusted, user-generated instructions influencing subsequent actions.