co-branding
Pass
Audited by Gen Agent Trust Hub on Feb 16, 2026
Risk Level: LOWNO_CODE
Full Analysis
- [NO_CODE] (SAFE): The skill consists entirely of Markdown instructions and contains no executable scripts, shell commands, or automation logic.
- [EXTERNAL_DOWNLOADS] (SAFE): No external dependencies, package managers (pip/npm), or remote scripts are referenced.
- [DATA_EXFILTRATION] (SAFE): The skill does not perform any network operations or access local sensitive files.
- [PROMPT_INJECTION] (SAFE): No patterns of instruction override, jailbreaking, or system prompt extraction were detected in the text or metadata.
- [INDIRECT PROMPT INJECTION] (INFO): The skill describes a workflow for processing 'partner-submitted collateral'. While this represents a theoretical ingestion point for untrusted data, the skill provides no automated capabilities (like code execution or file writing) that could be exploited by such data. This is a negligible risk for this specific implementation.
Audit Metadata