aicoin-trading
Fail
Audited by Gen Agent Trust Hub on Apr 8, 2026
Risk Level: HIGHCOMMAND_EXECUTIONDATA_EXFILTRATIONEXTERNAL_DOWNLOADS
Full Analysis
- [COMMAND_EXECUTION]: The script
scripts/trade.mjsusesexecSyncto execute commands by concatenating user-supplied arguments into a shell string. Although it attempts to wrap arguments in single quotes, this implementation is susceptible to command injection if arguments contain malicious shell characters. - [COMMAND_EXECUTION]:
scripts/exchange.mjsautomatically executesnpm installto download and install dependencies at runtime if theccxtpackage is missing, which involves executing arbitrary code from a remote registry. - [DATA_EXFILTRATION]: The skill accesses and reads sensitive
.envfiles from multiple locations, including the user's home directory (~/.openclaw/workspace/.envand~/.openclaw/.env), to harvest exchange API keys and secrets. - [COMMAND_EXECUTION]: The skill includes
scripts/auto-trade.mjswhich uses an internal environment flag (AICOIN_INTERNAL_CALL) to bypass the mandatory manual confirmation step defined in theSKILL.mdsafety rules, allowing for automated trade execution without user oversight. - [EXTERNAL_DOWNLOADS]: The skill fetches external code (the
ccxtlibrary) from the public NPM registry during execution if it is not already present in the local environment.
Recommendations
- AI detected serious security threats
Audit Metadata