research-frog

Automate iterative research workflows via a five-step loop with YAML config.

2|Updated Apr 30, 2026

Search Tags:#workflow #research #logs #orchestration #memory #iteration

Ready to run

npx skills add https://github.com/Siyou-Li/research-frog-skill --skill research-frog

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: research-frog
Source: https://github.com/Siyou-Li/research-frog-skill/tree/main/skills/research-frog
Command: npx skills add https://github.com/Siyou-Li/research-frog-skill --skill research-frog

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

Research projects often suffer from fragmented workflows, inconsistent memory across iterations, and manual, error-prone coordination. The Research Frog provides a disciplined, five-step loop that automates this process and preserves a reusable, verifiable history.

Core Features & Use Cases

Five-step iterative loop (review → execute → log → plan → archive) with persistent memory to support long-running projects.
Dispatches sub-skills (deep-research, academic-paper, academic-paper-reviewer, academic-pipeline) when a task requires literature scoping, drafting, reviewing, or pipeline orchestration.
YAML-config-driven adaptability that lets the same loop operate on different goals, hardware, and deliverables across multiple projects.

Quick Start

Initialize a project by filling out the YAML config and run the research-frog loop to start the first iteration.

Frequently Asked Questions about research-frog

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

FAQPage Schema

How do I automate iterative research workflows with persistent memory?▼

Automating iterative research workflows with persistent memory requires a structured five-step loop: review, execute, log, plan, and archive. This enforces tight log discipline and preserves a verifiable history for long-running projects.

What is the best way to coordinate literature review and drafting in a research project?▼

Coordinating literature review and drafting in a research project is best handled by dispatching sub-skills for literature scoping, drafting, and reviewing. This automates task routing while maintaining a unified orchestration loop.

Can I use a YAML config to adapt research orchestration for different project goals?▼

Yes, you can use a YAML config to adapt research orchestration for different project goals, hardware, and deliverables. The same loop operates across multiple projects by reading the configuration settings.

How does an automated research loop know when to stop iterating?▼

An automated research loop stops iterating when a clear termination condition is met. The orchestration loop evaluates this condition during the review and planning phases to ensure the research goal is achieved.

Why does my long-running research project lose context between iterations?▼

Long-running research projects lose context between iterations due to fragmented workflows and inconsistent memory. A disciplined loop with persistent logging and an archive step preserves a reusable history to prevent this.

Do I need to install external dependencies to run research orchestration in Claude Code?▼

No external dependencies are required to run research orchestration in Claude Code. The system operates autonomously using internal sub-skills and references without additional package installations.