nansen-holder-analysis

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

• Third-party content exposure detected (medium risk: 0.65). The required workflow runs nansen research ... CLI commands that fetch and return market/token holder/flow/trade data (public blockchain intelligence from an external source) as readable text into the agent context, creating an indirect prompt-injection risk from outsider-authored content.