Setup & Installation
Or with the ClawHub CLI, for registry-managed skill folders outside a full OpenClaw workspace:
What This Skill Does
Fetches recent papers from arXiv and Hugging Face, filters them by topic relevance using an LLM, and outputs summaries as JSON or via a local REST API. Configuration covers topics, time windows, arXiv categories, and model selection.
Combines source fetching, LLM-based relevance filtering, and structured JSON output in one pipeline, removing the need to wire those steps together manually.
When to use it
- Daily digest of new ML and NLP papers from arXiv
- Filtering AI research papers by custom topic list
- Feeding paper metadata into a downstream agent pipeline
- Running a local API server for polling-based paper workflows
- Scoping arXiv searches to specific categories like cs.AI or stat.ML
Example Workflow
Here's how your AI assistant might use this skill in practice.
User asks: fetch the last 24 hours of papers on LLM alignment and output JSON
- 1Bootstrap the repo with scripts/bootstrap.sh
- 2Create .env with OPENAI_API_KEY and set WINDOW_HOURS=24
- 3Update config/topics.json with an alignment topic entry
- 4Run scripts/run_cli.sh --window-hours 24 --sources arxiv,hf --json
- 5Report run_id, seen, kept counts from CLI output
JSON summary listing kept papers with titles, abstracts, and assigned topics
Requirements
Accounts, API keys, or tools you or your AI assistant may need to set up while using this skill.