kill-switch-binding
Pass
Audited by Gen Agent Trust Hub on Jun 16, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No malicious patterns or security vulnerabilities were detected. The skill provides legitimate administrative and safety controls for managing the operational state of the agent swarm.
- [COMMAND_EXECUTION]: The skill defines logic for handling specific Telegram commands (
/panic,/resume) and dashboard API requests. These actions are restricted to managing an internal status flag file (.harness/swarm/kill_switch.flag) and recording audit events to track administrative actions. - [DATA_EXFILTRATION]: No data exfiltration patterns were identified. Communication is limited to the configured Telegram operator chat and the internal dashboard for its intended administrative functions.
Audit Metadata