investor-research

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

• Third-party content exposure detected (high risk: 1.00). The SKILL.md explicitly directs the agent to call external search/scrape endpoints (e.g., orth api run exa /search, orth api run fiber /v1/investor-search and /v1/people-search, and the /v1/scrapes hint) that ingest public web/user-generated content which the agent is expected to read and use to drive researcher/outreach decisions, exposing it to untrusted third-party content that could inject instructions.