ml-paper-writing
Pass
Audited by Gen Agent Trust Hub on May 13, 2026
Risk Level: SAFE
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill references official conference domains (icml.cc, neurips.cc, aaai.org) and well-known academic repositories (github.com/acl-org, github.com/ICLR). These are trusted sources for paper templates and style files.
- [COMMAND_EXECUTION]: The NeurIPS template includes a Makefile for automating the LaTeX build process. It uses standard tools like
pdflatex,bibtex, andcurlto maintain style files from the official conference site. The operations are transparent and consistent with the skill's purpose. - [PROMPT_INJECTION]: Instructions within the skill, such as those prohibiting citation hallucination, are designed to ensure output quality and academic integrity. No attempts to bypass safety filters or override system constraints were found.
- [DATA_EXFILTRATION]: Network operations are limited to searching academic databases (Semantic Scholar, arXiv, CrossRef) via their official APIs for paper verification, which is the stated purpose of the citation workflow.
Audit Metadata