nansen-pm-deep-dive
Pass
Audited by Gen Agent Trust Hub on Mar 7, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill utilizes the 'nansen' CLI tool, which is a resource provided by the vendor 'nansen-ai'. These commands are used to query prediction market data such as order books and holder positions.
- [PROMPT_INJECTION]: The skill uses a template where an external variable (market_id) is interpolated into shell commands ($MID).
- Ingestion points: The 'market_id' variable, which is expected to be provided by a previous skill or the user.
- Boundary markers: None present; the variable is used directly in the shell script block.
- Capability inventory: Execution of shell commands via the 'nansen' CLI tool.
- Sanitization: There is no explicit sanitization or validation logic to ensure the market_id is strictly numeric before it is passed to the shell.
Audit Metadata