basket-query
Pass
Audited by Gen Agent Trust Hub on Apr 24, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: Utilizes the vara-wallet CLI and jq to fetch and parse blockchain information. These commands are limited to read-only queries as defined in the skill's scope.
- [PROMPT_INJECTION]: The skill processes data from external blockchain contracts, which serves as a surface for indirect prompt injection. Ingestion points: contract data such as names and descriptions returned by the vara-wallet call command. Boundary markers: none identified to isolate untrusted contract data. Capability inventory: blockchain interaction via the vara-wallet CLI. Sanitization: contract-derived content is not sanitized before incorporation into the agent's context.
Audit Metadata