task-status
Fail
Audited by Gen Agent Trust Hub on Mar 1, 2026
Risk Level: HIGHDATA_EXFILTRATIONCREDENTIALS_UNSAFECOMMAND_EXECUTION
Full Analysis
- [DATA_EXFILTRATION]: Hardcoded Telegram Target ID.
- The scripts
scripts/send_status.py,scripts/send_status_websocket.py, andscripts/send_status_with_logging.pyall include a hardcoded default value for theTELEGRAM_TARGETenvironment variable (7590912486). - This configuration causes the skill to exfiltrate status messages, task names, and potentially sensitive progress details to an external Telegram account by default whenever the
TELEGRAM_TARGETenvironment variable is not explicitly set by the user. - [CREDENTIALS_UNSAFE]: Hardcoded identifiers for external communication.
- The inclusion of a specific, non-placeholder Telegram ID used for data routing constitutes a security risk, as it redirects information flow to a destination not controlled by the end-user.
- [COMMAND_EXECUTION]: Subprocess execution of local CLI tools.
- The skill utilizes
subprocess.runandshutil.whichinscripts/send_status.pyandscripts/send_status_with_logging.pyto locate and execute theclawdbotcommand-line interface. - While this is a functional requirement for the skill's stated purpose, it executes commands using arguments derived from user input (task names and messages).
- [ENVIRONMENT_LEAKAGE]: Hardcoded user-specific file paths.
- Multiple files, including
SKILL.md,README.md, and the scripts, contain hardcoded absolute paths pointing toC:\Users\Luffy\.... scripts/send_status_with_logging.pyspecifically attempts to create and write logs toC:/Users/Luffy/clawd/logs, which will fail or expose environment details on systems with different user configurations.
Recommendations
- AI detected serious security threats
Audit Metadata