run-research-and-save-files

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

• Third-party content exposure detected (high risk: 0.90). The skill's orchestrator and required subagent briefs (SKILL.md plus references/industry/mission-briefs.md and references/industry/research-powerpack-and-explore.md) explicitly instruct subagents to run web searches/scrapes (smart-web-search, raw-web-search, smart-/raw-scrape-links, WebFetch) against public sites including Reddit/HN and to read those public, user-generated pages as part of the research workflow, so untrusted third-party content is ingested and can materially influence subsequent actions.