inno-idea-generation

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

• Third-party content exposure detected (high risk: 1.00). The skill’s SKILL.md research workflows and subagent coordination explicitly require using WebSearch/WebFetch to gather data from public sites (e.g., “Each agent uses WebSearch/WebFetch” in the Comprehensive Research Workflow and the Research Methods section), and scripts/research-sources.sh enumerates open/public sources like Reddit, G2, company websites and news — so the agent is expected to ingest untrusted, user-generated third‑party content that can influence decisions and next actions.