rw-research-novelty
Pass
Audited by Gen Agent Trust Hub on Jul 20, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill leverages official research methodologies from trusted organizations (NIH, NCBI, JBI, and GRADE) to guide its analysis. These external references are well-known academic standards and do not pose a security risk.
- [SAFE]: The provided script
scripts/self_check.pyis a static linter designed to verify the integrity and structure of the skill's own files. It does not perform network operations, execute external code, or access sensitive system paths. - [SAFE]: The instructions include a 'Source Discipline' (来源纪律) section that explicitly directs the agent to separate user-provided materials from its own methodology and publicly sourced data. This segregation is a best practice for mitigating indirect prompt injection risks when processing untrusted user input.
- [SAFE]: The skill explicitly forbids the generation of non-existent academic metadata (DOIs, papers, or datasets) and limits its scope to avoid accessing private user workspaces or external databases without permission.
Audit Metadata