hard-fix

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

• Third-party content exposure detected (high risk: 1.00). The skill's required Phase 2 explicitly launches a "research-online" agent (see Phase 2 in SKILL.md and the Research Agent prompt in references/templates.md) to fetch external solutions/known issues from the open web, and those findings are combined into the Phase 3 synthesis and used to drive fixes, so untrusted third‑party content can directly influence agent decisions.