shaman-jobs
Pass
Audited by Gen Agent Trust Hub on Mar 14, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No malicious behaviors or security vulnerabilities were identified in the analyzed files.
- [PROMPT_INJECTION]: The skill uses strong persona-driven instructions (e.g., 'Reality Distortion Field', 'Steve Jobs' digital soul'). While the tone is demanding and perfectionist, there are no attempts to override system constraints, extract system prompts, or disable safety filters.
- [DATA_EXFILTRATION]: There are no commands for network communication, sensitive file access, or credential harvesting. All operations are confined to text-based interaction within the AI session.
- [REMOTE_CODE_EXECUTION]: No external script downloads, package installations, or dynamic code execution patterns were detected.
- [INDIRECT_PROMPT_INJECTION]: The skill ingests user input via the
{{task}}placeholder. However, since the skill lack capabilities for system modification, network requests, or file writing, the risk associated with this injection surface is negligible.
Audit Metadata