The Agent Skills Directory

[SAFE]: The skill defines a rigorous scientific analysis workflow featuring evidence grounding and a self-refinement loop specifically designed to minimize hallucinations and ensure factual accuracy when processing external academic papers.
[SAFE]: The Python script scripts/evaluate_smoke_results.py is a benign utility for processing local JSON test logs using standard libraries; it does not perform any network operations, command execution, or access sensitive system files.
[SAFE]: The skill identifies ArXiv as its primary data source, which is a well-known and trusted repository for research papers. All instructions are focused on academic summarization and verification without attempting to bypass safety filters or exfiltrate data.

paper-summary