research-init

Initialize a research topic scaffold with seeds and metadata in Research Harness.

Updated Aug 27, 2026
One-click install
npx skills add https://github.com/Biajin-PKU/research-harness --skill research-init
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: research-init
Source: https://github.com/Biajin-PKU/research-harness/tree/main/codex-skills/research-init
Command: npx skills add https://github.com/Biajin-PKU/research-harness --skill research-init

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This skill streamlines the process of starting a new research topic or project integration by creating a minimal, guided scaffold within Research Harness, reducing setup time and ensuring consistent guidance.

Core Features & Use Cases

  • Topic scaffolding: creates a new topic with a slug, repository metadata, and initial guidance scaffolds.
  • Guidance integration: updates AGENTS.md and CLAUDE.md when requested to align with the topic.
  • Seed ingestion: ingests provided seed papers or references to seed the workspace and provenance.
  • Audit-ready setup: records assumptions for missing metadata to maintain traceability across sessions.

Quick Start

Provide a topic name and optional seeds to scaffold a new research topic.

Frequently Asked Questions about research-init

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

FAQPage Schema
How do I initialize a new research topic and scaffold a workspace?

To initialize a new research topic, provide a topic name and optional seed papers. The system then derives a slug, ingests seeds, establishes a minimal scaffold, and updates AGENTS.md and CLAUDE.md guidance when requested.

What is topic scaffolding and how does it help start a research project?

Topic scaffolding establishes a minimal, guided structure for a new research project by generating a topic slug, repository metadata, and initial guidance files, reducing setup time and ensuring consistent agent alignment across sessions.

How do I ingest seed papers when bootstrapping a research workspace?

You ingest seed papers by providing them alongside your topic name during initialization. The workspace uses these references to seed initial provenance and establish the foundational context for your research topic.

Does initializing a research topic automatically update CLAUDE.md and AGENTS.md?

Updating AGENTS.md and CLAUDE.md occurs when explicitly requested during the topic integration process. This guidance integration aligns the workspace configuration with your newly established research topic scaffold.

What happens to missing metadata when I set up a new research project scaffold?

When missing metadata occurs during setup, the system records assumptions to maintain traceability across sessions. This audit-ready approach ensures your research topic initialization remains fully documented without blocking the scaffold creation.

Do I need to provide a topic slug before bootstrapping a new research topic?

You do not need to provide a topic slug before bootstrapping. The initialization process automatically derives the topic slug from your provided topic name if one is not explicitly supplied during setup.