super-pm

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

• Third-party content exposure detected (high risk: 1.00). The skill explicitly uses WebSearch and Agent-dispatched subagents in multiple required workflows (e.g., pm-search, pm-market, and pm-brainstorm SKILL.md) to fetch and integrate results from public sites and social platforms (36kr.com, huxiu.com, Weibo, Zhihu, ProductHunt, etc.), and those external findings are parsed by the agent to produce reports and drive recommendations/actions—exposing it to untrusted third‑party content that could carry indirect prompt injection.