langgraph

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

• Third-party content exposure detected (high risk: 0.80). The skill's patterns (references/patterns.md, "Basic Agent Graph") include a search tool described as "Search the web for information" which is bound into the agent's tools/ToolNode and thus indicates the agent will fetch and consume open/public web content returned by that tool, exposing it to untrusted third-party input.