effect-ai-tool

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

• Third-party content exposure detected (high risk: 0.90). The skill explicitly includes a FetchTool example that performs fetch(url) on arbitrary URLs (Common Patterns -> "FetchTool") and documents a providerDefined WebSearch tool (Provider-Defined Tools -> "openai.web_search") whose results are returned to the LLM (Tool Result Flow), meaning the agent ingests untrusted public web content that can influence subsequent actions.