research

Orchestrate multi-stage research workflows with resumable progress and asynchronous experiment execution.

1.6k|208|Updated Apr 9, 2026
One-click install
npx skills add https://github.com/skyllwt/AutoSci --skill research-skyllwt
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: research
Source: https://github.com/skyllwt/AutoSci/tree/main/.claude/skills/research
Command: npx skills add https://github.com/skyllwt/AutoSci --skill research-skyllwt

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

It eliminates the manual, error-prone work of planning and running an entire research cycle from idea generation through experiments to a paper-ready output.

Core Features & Use Cases

  • End-to-end orchestration: coordinates idea discovery, experiment design, deployment, result collection, and verdict-driven iteration.
  • Pipeline recovery & async execution: saves state to wiki/outputs/pipeline-progress.md and supports non-blocking experiment deployment and later collection.
  • Paper production flow: generates a paper plan, drafts, refines, and compiles—optionally skipping paper writing while still evaluating experiments.

Quick Start

Run the full pipeline by providing a one-sentence research direction and enabling auto mode: /research "sparse LoRA for edge devices" --auto --venue ICLR

Frequently Asked Questions about research

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I automate a full research pipeline from idea generation to paper writing?

A research pipeline coordinates end-to-end stages from idea discovery through experiment execution to paper-ready output. It orchestrates sub-skills for ideation, experiment design, deployment, and evaluation, while persisting progress to enable session recovery.

Can I resume an interrupted research workflow without losing my experiment progress?

You can resume an interrupted research workflow by utilizing built-in state recovery. The pipeline saves progress to a persistent markdown file, allowing you to restart asynchronous experiment collection without losing prior stage gating results.

What is the best way to run non-blocking experiment deployment during a research cycle?

Non-blocking experiment deployment is handled through asynchronous execution within pipeline orchestration. You can deploy experiments and collect results later, enabling coordinated multi-stage workflows without halting overall research progress.

Does the pipeline support skipping paper writing while still evaluating experiment results?

The pipeline supports skipping paper writing while still evaluating experiment results. You can run orchestration to generate experiment verdicts and iterate, optionally excluding paper drafting and refinement stages from the workflow.

How do I execute a complete research workflow using only a single prompt?

You can execute a complete research workflow by providing a one-sentence research direction and enabling auto mode. This triggers the pipeline to autonomously coordinate idea discovery, experiment execution, and paper compilation.