matematic-expert-panel
Pass
Audited by Gen Agent Trust Hub on Jul 14, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill serves as a methodology and documentation guide for business workshops. It does not contain executable scripts or active code blocks.
- [PROMPT_INJECTION]: While the skill utilizes role-play and personas, these are strictly defined as business analysis perspectives (e.g., 'Compliance Officer', 'IT Security Lead') aimed at identifying organizational risks. There are no attempts to bypass safety filters or override system instructions.
- [REMOTE_CODE_EXECUTION]: The skill mentions the 'SuperClaude_Framework' as an architectural reference for its methodology. It explicitly states that it is a 'cherry-pick' of the pattern and not a full installation, and it contains no commands to download or execute remote code.
- [DATA_EXFILTRATION]: The skill is configured with 'data-residency: local' and 'pii-egress: none'. No network-enabled tools or unauthorized data access patterns were identified in the instructions.
- [EXTERNAL_DOWNLOADS]: References to external GitHub repositories are for methodology attribution and source credits only. No automated download or dependency fetching is performed by the skill.
Audit Metadata