tw-research-grant-writer

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

• Third-party content exposure detected (high risk: 0.90). The skill's required workflow in Step 3 explicitly runs "WebSearch：搜尋近 5 年最具影響力的相關研究" (SKILL.md Step 3, AI 執行), which fetches and ingests open/public research content that the agent is expected to read and use to shape the proposal text and citations, exposing it to untrusted third‑party input that could influence subsequent actions.