tp

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

• Third-party content exposure detected (high risk: 0.80). The skill's flow (references/flow.md — Execution Flow step 1 and the "Research-driven" principle) explicitly instructs the agent to use WebSearch/WebFetch to retrieve web content (official docs/best practices) from the public internet, which the agent will read and use to influence plan decisions, exposing it to untrusted third-party content.