ce-deepen-spec

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

• Third-party content exposure detected (high risk: 0.80). The skill's Workflow Phase 3 explicitly calls external research lanes (e.g., best-practices-researcher: "max 5 external sources" and framework-docs-researcher) and states those retrieved external URLs/dates are used as grounding evidence for deepening the spec, meaning untrusted third-party web content would be read, interpreted, and could materially change the agent's decisions.