dag-parallel-executor

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

• Third-party content exposure detected (medium risk: 0.65). The skill’s runtime workflow spawns subagents (e.g., Task(... subagent_type="web-researcher") in the worked example) that can fetch public web content, and that fetched free-form page text would be ingested into the agent LLM context via the Task tool’s prompt/context.