core-fix-skill-docs

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

• Third-party content exposure detected (high risk: 0.85). The skill’s runtime path can fetch outsider-authored free text from public web content (docs.rs pages) via agent-browser open/get text or WebFetch, then writes that fetched prose into references/*.md which is subsequently read/used by the agent—an indirect prompt-injection exposure channel.