customaize-agent:thought-based-reasoning

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

• Third-party content exposure detected (high risk: 1.00). The skill's ReAct section explicitly instructs the agent to perform Search/Lookup actions on public web queries (e.g., "Action 1: Search[Colorado orogeny]" and subsequent Search[High Plains]), ingesting observations from open third-party sources that the agent then uses to drive reasoning and follow-up actions, which exposes it to untrusted user-generated/public content and potential indirect prompt injection.