ml-paper-writing

Pass

Audited by Gen Agent Trust Hub on Sep 15, 2026

Risk Level: SAFE
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill includes a Makefile for NeurIPS 2025 that uses curl to download style files directly from the official conference domain (media.neurips.cc). This is a legitimate utility for maintaining up-to-date templates.
  • [DATA_EXFILTRATION]: The skill uses Python snippets to interact with well-known academic APIs, including Semantic Scholar, CrossRef, and arXiv. These operations are restricted to citation metadata retrieval and are integral to the skill's primary purpose of academic verification.
  • [REMOTE_CODE_EXECUTION]: The skill recommends the use of 'Exa MCP' for real-time academic search. The installation instructions provided for Claude Code and VS Code point to the official exa.ai domain and use standard MCP tool configuration patterns.
  • [COMMAND_EXECUTION]: The instructions include routine shell commands (e.g., ls, find, grep, cp) and LaTeX compilation commands (pdflatex, bibtex, latexmk). These are standard for exploring research repositories and building documents.
  • [NO_CODE]: The skill primarily consists of configuration files, LaTeX templates, and documentation. The executable snippets provided are for instructional purposes or environment setup (e.g., pip install semanticscholar).
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 15, 2026, 10:30 AM