writing-agents

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

• Third-party content exposure detected (medium risk: 0.65). The required runtime workflow is “Writing Agents” (research via WebSearch, explore codebase, then write agent files), and the WebSearch step can fetch public web pages whose readable text may be ingested into the agent’s LLM context, creating an outsider free-text injection surface.