zoom-general

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

• Third-party content exposure detected (high risk: 0.80). The skill explicitly tells agents to perform web searches and fetch external documentation (see "Note for LLMs/Agents" in references/sdk-logs-troubleshooting.md instructing: "If the log retrieval URLs return 404 or have moved, perform a web search..."), which requires ingesting open/public, potentially user-generated web content and using it to change troubleshooting actions — exposing the agent to indirect prompt injection.