liquidity-planner
Pass
Audited by Gen Agent Trust Hub on Sep 20, 2026
Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATIONINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADSDYNAMIC_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill utilizes command-line tools such as
curl,jq, andcastto fetch pool data and verify token contracts on-chain. It also executes a locally provided Node.js script to aggregate APR and reward data. - [DATA_EXFILTRATION]: Performs an initialization ping to the vendor domain
pancakeswap.ai, transmitting basic system metadata including operating system and architecture for telemetry. - [INDIRECT_PROMPT_INJECTION]: The skill is susceptible to indirect injection through its data ingestion points, including DexScreener, DefiLlama, and the PancakeSwap Explorer API. Boundary markers include explicit security warnings and the use of the
--data-urlencodeflag in scripts. The capability inventory includescurl,node, andcastexecution. Sanitization is implemented via strict address regex validation and safe shell variable interpolation usingjq. - [EXTERNAL_DOWNLOADS]: Retrieves real-time token and pool information from well-known services and official PancakeSwap API endpoints.
- [DYNAMIC_EXECUTION]: Runs a distributed script (
discover-pools.mjs) to process reward logic and performs on-chain interactions using thecastutility.
Audit Metadata