peeps

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

• Third-party content exposure detected (high risk: 0.80). The SKILL.md explicitly instructs the agent to "always search the web for the person (name + any context provided)" and to look up LinkedIn, personal/company websites and Wikipedia (and even fetch SKILL.md updates from raw.githubusercontent.com), meaning it fetches and reads untrusted public web content to pre-fill profiles and drive follow-up actions.