multi-agent-creator
Pass
Audited by Gen Agent Trust Hub on Feb 20, 2026
Risk Level: SAFEPROMPT_INJECTIONNO_CODE
Full Analysis
- PROMPT_INJECTION (LOW): The file
description_ar.txtcontains self-referential phrases ('لا داعي للقلق بشأن هذه المشكلة' and 'لا يوجد أي مشكلة في هذا الأمر') which translate to 'No need to worry about this problem' and 'There is no problem in this matter.' This matches the Category 8e pattern where content attempts to pre-emptively declare its own safety to the auditor or agent. Additionally, the skill has an Indirect Prompt Injection surface (Category 8): 1. Ingestion points: User-provided business process descriptions (Step 1) used to generate agent templates. 2. Boundary markers: Absent in the provided skill body. 3. Capability inventory: Task orchestration using theTodoWriteandTasktools. 4. Sanitization: None mentioned for user inputs interpolated into templates. - NO_CODE (SAFE): The skill contains no executable scripts, binaries, or dependency configuration files. It consists entirely of markdown and text files.
Audit Metadata