omcustom-loop
Pass
Audited by Gen Agent Trust Hub on May 15, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill documentation defines a session management workflow using built-in platform features.
- [PROMPT_INJECTION]: Utilizes prompt hooks to provide instructions to the orchestrator for continuation. The implementation includes safety boundaries, such as a maximum of three auto-continues, which ensures human-in-the-loop oversight is maintained.
- [COMMAND_EXECUTION]: References a shell script titled
task-outcome-recorder.shwithin the documentation for hooks configuration; however, no script files are included in the skill distribution.
Audit Metadata