get-context

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

• Third-party content exposure detected (high risk: 0.70). The required workflow launches a code-context:context-researcher agent, and that agent’s “code context” research typically involves fetching/parsing external repository/web content (outsider-authored free text) at runtime and injecting it into the LLM context; this skill itself provides no constraints preventing that.