Setup & Installation
Or with the ClawHub CLI, for registry-managed skill folders outside a full OpenClaw workspace:
Version History
- Updated workflow instructions to prioritize proactive user interaction and guidance. - Clarified that the user must first open and read the configuration from the downloaded GitHub repo. - Emphasized assisting the user in providing preferences for topics, time window, and core parameters. - Added a note to default to 24 hours for the time window unless the user specifies otherwise. - Added recommendation to default OpenAI model to "gpt-5-mini" for optimal tradeoff. - Added reminders for skill developers to help the user tune for good results. - Minor edits for clarity, directness, and improved step-by-step guidance.
What This Skill Does
Fetches recent papers from arXiv and Hugging Face, filters them by topic using an LLM relevance classifier, and produces summaries. Runs as a one-off CLI command or a persistent local API server. Results are stored in a SQLite database.
Combines source fetching, LLM-based relevance filtering, and summarization in a single configurable pipeline instead of requiring separate tools for each step.
When to use it
- Daily digest of new ML/AI papers filtered to your research topics
- Feeding filtered paper metadata into a downstream agent pipeline
- Polling arXiv overnight and saving results for async review
- Tracking papers in specific arXiv categories over a custom time window
- Running a local API so a UI or cron job can query recent papers
Example Workflow
Here's how your AI assistant might use this skill in practice.
User asks: Get me today's papers on LLM safety from arXiv
- 1Bootstrap the repo using the provided bootstrap script
- 2Set OPENAI_API_KEY in the .env file
- 3Update config/topics.json with an LLM safety topic entry including relevant keywords
- 4Run the CLI with --window-hours 24 --sources arxiv
- 5Read the JSON output for run_id, seen, kept counts and query /api/papers for summaries
A filtered list of arXiv papers matching the LLM safety topic, each with title, abstract summary, and metadata stored in papers.sqlite3
Requirements
Accounts, API keys, or tools you or your AI assistant may need to set up while using this skill.