AutoSci
AI research agent that reads, experiments, and writes papers
All Skills in This Repository (47)
Pure Emerald Level Indicatorsresearch
Orchestrate multi-stage research workflows with resumable progress and asynchronous experiment execution.
/daily-arxiv
Generate daily arXiv paper digests matched to wiki interests.
/init
Orchestrate ΩmegaWiki initialization and parallel paper ingestion from raw sources.
exp-design
Design end-to-end experiment suites with ablation-driven iterative refinement for wiki-based research workflows.
paper-draft
Generate a compilable LaTeX paper from PAPER_PLAN and wiki pages.
/discover
Rank candidate research papers from anchor, topic, wiki, or venue inputs.
ask
Retrieve and synthesize wiki pages into cited answers for research questions.
/setup
Configure missing Semantic Scholar, DeepXiv, and Review LLM API keys in `.env`.
/exp-pilot-run
Execute pilot experiments from Pilot Spec YAML and return raw metrics.
/exp-status
Check local screen sessions and remote endpoints for experiment status anomalies.
visualize
Generate Obsidian graph configurations and Canvas knowledge maps from OmegaWiki JSONL data.
/paper-compile
Compile LaTeX papers with latexmk and generate a submission checklist.
Frequently Asked Questions
FAQPage SchemaHow to install AutoSci?▼
Run `npx skills add skyllwt/AutoSci --all -g -y` in your terminal to install all skills in this suite globally.
What does AutoSci do for researchers?▼
It runs the full research pipeline: ingesting papers into a knowledge wiki, generating and scoring ideas, designing and executing experiments, and drafting submittable LaTeX papers.
Can AutoSci recommend new papers to read?▼
Yes. The daily-arxiv and discover skills rank fresh arXiv papers and venue proceedings against your existing wiki interests and deliver a personalized shortlist.
Does AutoSci work with Claude Code and Codex?▼
Yes. The main branch targets Claude Code, and separate preview branches provide Codex and OpenCode adaptations, all driven by the same bilingual skill source.
Do I need coding skills to use AutoSci?▼
No. You describe your research direction in plain language, and the agent handles paper ingestion, experiment code, and paper writing, asking for your confirmation at key gates.
Related Repositories in Education & Research
View All in Education & Research→learn-claude-code
Learn to build AI agent harnesses from scratch
ai-agents-for-beginners
Learn to build, deploy, and secure AI agents step by step
33-js-concepts
Master 33 essential JavaScript concepts with tested, verified guides