univer-pro-integrate

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

• Third-party content exposure detected (high risk: 0.80). The skill clearly ingests untrusted external content: references/exchange-guide.md and references/facade-extension-pro.md show univerAPI.importXLSXToUnitIdAsync(fileOrUrl) with examples importing from remote URLs (e.g., 'https://example.com/data.xlsx'), and references/collaboration-guide.md shows loading server units and fetch calls to arbitrary server endpoints, so the agent will read/interpret third‑party content that can influence actions.