DeepScientist
Autonomous local research studio from baselines to paper drafts
All Skills in This Repository (21)
Pure Emerald Level Indicatorswrite
Generate publication-ready paper drafts from validated quest evidence.
optimize
Shape and govern end-to-end optimization workflows for algorithm-first quests.
baseline
Standardize a reference object for quest evaluation with plan and checklist requirements.
analysis-campaign
Coordinate code-based follow-up analyses with campaign plans and per-slice runs.
scout
Clarify quest framing by outlining objectives, data requirements, and evaluation criteria.
idea
Convert a baseline and problem frame into literature-grounded research directions.
review
Audits drafts and papers for evidence-grounded critique and produces a revision plan.
figure-polish
Polish research figures using a render-inspect-revise workflow and Matplotlib style.
alphaxiv-paper-lookup
Fetch structured AI-generated overviews of arXiv papers from alphaxiv.org.
experiment
Plan, execute, and log auditable quest experiments with PLAN.md and CHECKLIST.md.
decision
Creates canonical records for quest decisions with verdicts, actions, and evidence paths.
finalize
Consolidate final claims, limitations, recommendations, and graph exports for quest closure.
Frequently Asked Questions
FAQPage SchemaHow to install DeepScientist?▼
Run `npx skills add ResearAI/DeepScientist --all -g -y` in your terminal to install the full skill suite globally.
What does DeepScientist actually do?▼
It is a local-first autonomous research studio that takes a paper, repo, or research question and drives it through baselines, experiments, analysis, and paper drafts on your own machine.
Can DeepScientist reproduce a baseline from a paper?▼
Yes. The baseline skill attaches, imports, or reproduces a comparator, verifies its metrics, and records a durable comparison contract before any experiment starts.
Which AI agents work with DeepScientist?▼
It ships built-in runners for Codex, Claude Code, Kimi Code, and OpenCode, and all skills follow the standard SKILL.md format.
Can I take over from the AI during a research quest?▼
Yes. You can pause, edit plans or code, inspect every artifact in the web workspace or TUI, and hand control back to the agent at any time.
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