prd-to-plan

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

• Third-party content exposure detected (high risk: 0.80). The skill explicitly instructs the agent to "read or fetch the PRD from wherever it lives" (e.g., GitHub wiki, Notion, Confluence) and to pull external context from user-provided Linear, Figma, and Notion URLs, meaning it will ingest and act on arbitrary third-party/user-provided content as part of its workflow.