skill-create

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

• Third-party content exposure detected (high risk: 0.80). The skill's PHASE 1 Step 2 "Web Research (Token-Optimized)" in SKILL.md explicitly instructs the agent to use a Task/Explore subagent to search public web sources and return "Sources used" and research that will be incorporated into questions, quality checks, templates, and generation decisions, meaning untrusted third‑party content is fetched and directly influences the agent's actions.