research-session

Load research state files and generate a briefing with sub-skill routing.

43|2|Updated Apr 10, 2026
One-click install
npx skills add https://github.com/andrehuang/researcher-pack --skill research-session
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: research-session
Source: https://github.com/andrehuang/researcher-pack/tree/main/skills/research-session
Command: npx skills add https://github.com/andrehuang/researcher-pack --skill research-session

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Structured research work often stalls at session start—researchers must gather context, decide next steps, and route to the right tools. This Skill acts as a meta-orchestrator to load state, generate a compact briefing, and hand off to the appropriate sub-skills.

Core Features & Use Cases

  • Context loading from key files including .claude/research-state.yaml, events.jsonl, IDEAS.md, and CLAUDE.md to surface current work and backlog.
  • Briefing generation that summarizes training status, knowledge base health, active threads, and backlog with recommended focus.
  • Automatic routing to sub-skills such as /paper-read, /research-companion, /academic, /mental-gym, /weekly-review, or /orchestrate as tasks evolve.
  • Phase-based workflow that guides a user from Context Load through BRIEFING to ROUTE and TRANSITION, with options for "Briefing only" or "Deep work".
  • On-demand modes to support quick checks or deep work sessions.

Quick Start

Say "research-session" to load the current state, generate a briefing, and route to the appropriate sub-skills.

Frequently Asked Questions about research-session

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

FAQPage Schema
How do I bootstrap a structured research session and generate a briefing?

To bootstrap a research session, load context from local state files like research-state.yaml, events.jsonl, and IDEAS.md, then generate a concise briefing that summarizes current threads, backlog, and recommended focus.

How do I start a research workflow by loading state from local markdown and yaml files?

Triggering this research workflow loads state from local files including research-state.yaml, IDEAS.md, CLAUDE.md, and wiki pages, then generates a compact briefing summarizing knowledge base health and active threads.

Can I route my research session to specific sub-skills like paper-read automatically?

Yes, this orchestrator automatically routes your research session to sub-skills such as paper-read, research-companion, academic, and mental-gym as your research goals evolve throughout the workflow phases.

What files do I need to set up before orchestrating a research session?

You need local research state files including research-state.yaml, events.jsonl, IDEAS.md, CLAUDE.md, and wiki pages with defined frontmatter and allowed-tools to drive the workflow orchestration.

Does this research session orchestrator support quick briefing checks without deep work?

Yes, the research session orchestrator supports on-demand modes including a quick 'Briefing only' check and a 'Deep work' session to support varying research needs.

What is the best way to manage research context and decide next steps automatically?

Using a meta-orchestrator to load state, generate a compact briefing, and hand off to sub-skills is the best way to manage research context, as it eliminates the manual effort of gathering context and routing tools.