estimation-fermi

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

• Third-party content exposure detected (high risk: 0.80). This skill's workflow and methodology explicitly instruct the agent to consult and validate estimates using open/public third‑party sources (e.g., "Quick reliable sources: Google, Wikipedia, Public company filings" in resources/methodology.md and "After Fermi estimate, quickly check if public data exists" in SKILL.md), which requires ingesting untrusted web content that could influence subsequent decisions and tool use.