aicoin-freqtrade
Fail
Audited by Gen Agent Trust Hub on May 6, 2026
Risk Level: HIGHCREDENTIALS_UNSAFEEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
- [CREDENTIALS_UNSAFE]: The file
lib/defaults.jsoncontains a hardcodedaccessKeyIdandaccessSecret. While the comments suggest these are public free-tier credentials, hardcoding secrets within the skill's source files is a high-risk practice.\n- [COMMAND_EXECUTION]: The scriptsscripts/ft-deploy.mjsandscripts/ft.mjsutilizeexecSyncto perform system-level operations. These include managing Freqtrade daemon processes viasupervisorctlandpgrep, checking for system dependencies, and executing thefreqtradebinary.\n- [REMOTE_CODE_EXECUTION]: The scriptscripts/ft-deploy.mjsimplements remote code execution by downloading and running theuvinstaller fromastral.shusing acurl | shpipe. It also clones the Freqtrade repository from GitHub and executes asetup.shscript. Furthermore,lib/strategy-builder.mjsdynamically generates Python code for trading strategies, which is then written to the file system and executed.\n- [EXTERNAL_DOWNLOADS]: The skill downloads historical data and market signals from external sources such asopen.aicoin.comand GitHub repositories.\n- [DATA_EXFILTRATION]: The skill accesses sensitive local files including.envand.ft_api_passto retrieve trading credentials. The combination of reading sensitive local data and having functional network clients creates a potential surface for data exfiltration.\n- [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection from untrusted external data:\n - Ingestion points: Market analysis, AI predictions, and news data are fetched via the AiCoin API in
lib/aicoin-api.mjs.\n - Boundary markers: External data is processed without delimiters or instructions to the agent to ignore embedded commands.\n
- Capability inventory: The skill has powerful system capabilities, including file system write access and the ability to execute shell commands through
execSync.\n - Sanitization: There is no evidence of filtering or sanitizing text data received from the remote API before it is handled by the agent.
Recommendations
- AI detected serious security threats
Audit Metadata