farming-planner

Fail

Audited by Gen Agent Trust Hub on Nov 8, 2026

Risk Level: HIGHREMOTE_CODE_EXECUTIONDYNAMIC_EXECUTIONDATA_EXFILTRATIONINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [REMOTE_CODE_EXECUTION]: The Python reference scripts (fetch-farms.py and fetch-syrup-pools.py) dynamically install the 'requests' library at runtime if not present. This is done via unpinned 'pip install' calls through subprocess.check_call and os.system, which exposes the system to potential supply chain attacks or dependency confusion.
  • [DATA_EXFILTRATION]: In Step 0 (Initialize) of SKILL.md, a curl command is used to ping a vendor-owned telemetry endpoint (pancakeswap.ai), transmitting environment metadata such as operating system and CPU architecture.
  • [INDIRECT_PROMPT_INJECTION]: The skill reads from several external APIs (Explorer, DexScreener, CoinGecko) whose content (token names, symbols) is potentially attacker-controlled. This creates a surface for prompt injection, though the skill provides explicit instructions to the agent to treat this data as untrusted.
  • [COMMAND_EXECUTION]: The skill generates shell commands using 'cast' and 'curl' based on user inputs. While it mandates regex-based input validation and quoting, any bypass or implementation error by the agent could lead to command injection.
Recommendations
  • HIGH: Downloads and executes remote code from: https://explorer.pancakeswap.com/api/cached/pools/list?orderBy=volumeUSD24h&protocols=v2&protocols=v3&protocols=stable&protocols=infinityBin&protocols=infinityCl&protocols=infinityStable&chains=bsc&chains=ethereum&chains=base&chains=arbitrum&chains=zksync&limit=100, https://sol-explorer.pancakeswap.com/api/cached/v1/pools/list?orderBy=volumeUSD24h&protocols=v3&limit=100 - DO NOT USE without thorough review
Audit Metadata
Risk Level
HIGH
Analyzed
Nov 8, 2026, 08:21 PM