kimu-viewpoint-script
Pass
Audited by Gen Agent Trust Hub on Sep 9, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No malicious patterns such as prompt injection, data exfiltration, or obfuscation were detected across the skill files.
- [COMMAND_EXECUTION]: The skill uses a local script
scripts/validate_viewpoint_script.pyto validate draft formatting. Analysis of the script shows it is restricted to basic text processing, regex matching, and frequency counting. It does not perform network operations, access sensitive files, or execute arbitrary code. - [INDIRECT_PROMPT_INJECTION]: While the skill processes external course materials, the risk is minimized by strict output formatting requirements and the use of a validation script that enforces specific structural constraints (
【本段P数字】).
Audit Metadata