xukp20 skills on Remote OpenClaw

arxiv-paper-processor

Downloads source files and PDFs from ArXiv, then has the model read the full paper and write a structured summary in a chosen language. Supports single-paper and batch modes. Scripts handle only artifact retrieval; all reading and synthesis is model-driven.

arxiv-search-collector

A four-step workflow for collecting arXiv papers where the model designs search queries, fetches metadata via the arXiv API, filters results by relevance, and merges selected papers into a structured directory. Query planning and relevance decisions are made by the model, not hardcoded rules.

arxiv-summarizer-orchestrator

Orchestrates a three-stage pipeline for collecting, summarizing, and reporting on arXiv papers by chaining three sub-skills: arxiv-search-collector, arxiv-paper-processor, and arxiv-batch-reporter. Supports scheduled execution with configurable parallel or serial per-paper processing and manual language selection across all generated outputs.

arxiv-batch-reporter

Assembles a final collection report from multiple arXiv paper summaries in two steps. A model writes a structured Markdown template with per-paper placeholders, then a script injects each paper's verbatim brief conclusion and abstract URL from existing summary.md files. Designed to run after per-paper summaries already exist.