decision-tree
Pass
Audited by Gen Agent Trust Hub on Aug 22, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [NO_CODE]: The skill consists entirely of Markdown instructions and historical/conceptual examples. It does not include scripts (Python, Node.js, Shell) or configuration files that could execute commands.
- [COMMAND_EXECUTION]: There are no shell commands, subprocess calls, or system interactions present in the skill.
- [PROMPT_INJECTION]: The instructions for the AI are pedagogical and structured. It includes 'WAIT' markers for step-by-step coaching, but these are for managing the user interaction flow rather than bypassing safety filters.
- [EXTERNAL_DOWNLOADS]: The skill references external URLs (deciqai.com, github.com) for documentation and versioning, but these are informational links and not used for fetching code or runtime assets. These references involve the vendor's own infrastructure (deciqai).
- [DATA_EXFILTRATION]: No mechanisms for reading sensitive files or sending data to external endpoints were detected.
Audit Metadata