clawby-data
Fail
Audited by Gen Agent Trust Hub on Nov 2, 2026
Risk Level: CRITICALEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill manages the installation of several external tools and libraries. \n
- Fetches the official binance-cli from Binance's GitHub repository. \n
- Installs the akshare Python package from PyPI for China market data. \n
- Installs several trading connectors, including @bitget-ai/bitget-agent-cli, @okx_ai/okx-trade-cli, and @binance/agentic-wallet from the npm registry. \n
- All identified download sources are official repositories or registries of trusted organizations. \n- [COMMAND_EXECUTION]: The skill's primary functionality relies on executing external command-line interfaces to interact with trading venues and data sources. \n
- Leverages CLI tools like okx, bgc, binance-cli, and baw to perform on-chain and exchange-side operations. \n
- Runs local scripts to manage automated trading logic, such as the daemon.mjs volume farming loop and ak.py for localized data extraction. \n- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user-generated content from external sources, creating an indirect prompt injection surface. \n
- Ingests data such as market questions and outcome names from prediction markets like 42.space and Polymarket. \n
- The instructions explicitly identify this risk and provide clear guidance to the agent to treat these fields as untrusted and avoid following any instructions contained within them.
Recommendations
- Contains 1 malicious URL(s) - DO NOT USE
Audit Metadata