| name | nature-literature-pipeline |
| description | Build a repeatable literature discovery, scoring, reading, and archiving pipeline for RL, MARL, LLM, and LLM Agent research. Use for daily/weekly paper tracking, topic watchlists, benchmark monitoring, and research-note automation. |
| version | 2.0.0-agent-rl |
Literature pipeline skill
Workflow
- Define scope: topic, venues, keywords, authors, benchmarks, arXiv categories.
- Define sources: arXiv, OpenReview, Papers with Code, venue proceedings, official benchmark repos.
- Score papers using six axes: relevance, novelty, evidence, benchmark match, reproducibility, idea value.
- Produce a triage table: read now, skim, archive, ignore.
- Generate one-page notes for high-priority papers.
- Maintain a local index with BibTeX, code links and relation to current projects.
Output format
Pipeline design:
Sources:
Search queries:
Scoring rubric:
Weekly triage table:
Archive structure:
Red lines
Do not present unverified paper metadata as final. Do not automate downloads from restricted sources.