Yeachan-Heo/My-Jogyo
Your one-click scientific research lab for Opencode - with seamless .ipynb and REPL integration
3 Releases
Latest: 5mo ago
v0.5.1 - Bug Fixes for MCP Serverv0.5.1Latest
๐ฆ MCP Server Build Error (#6)
- Fixed `Could not resolve "zod"` error when building the MCP server by adding `nodePaths` to esbuild config.
๐ฆ Hardcoded Bridge Path (#5)
- Fixed hardcoded bridge path in `python-repl.ts` that broke MCP server for cloned repos. The new implementation:
- Supports `GYOSHU_BRIDGE_PATH` environment variable override
- Falls back through multiple candidate paths for different installation scenarios
- Provides helpful error messages when bridge script cannot be found
๐ฆ Installation
- ```bash
- cd src/mcp
- npm install
- npm run build
- ```
- Register with Claude Code:
- ```bash
- claude mcp add gyoshu-mcp "$(pwd)/src/mcp/build/index.cjs"
- + 1 more
๐ Full Changelog
- https://github.com/Yeachan-Heo/My-Jogyo/compare/v0.5.0...v0.5.1
v0.5.0 - Claude Code MCP Supportv0.5.0
๐ฆ Claude Code MCP Server
- This release adds full Model Context Protocol (MCP) support for Claude Code integration.
๐ฆ Installation
- ```bash
- git clone https://github.com/Yeachan-Heo/My-Jogyo.git
- cd My-Jogyo/src/mcp
- npm install && npm run build
- claude mcp add gyoshu-mcp "$(pwd)/build/index.cjs"
- ```
โจ Features
- 12 Research Tools exposed via MCP:
- `python_repl` - Execute Python code with marker detection
- `research_manager` - Create/manage research sessions
- `session_manager` - Session lifecycle management
- `notebook_writer` - Jupyter notebook operations
- `gyoshu_completion` - Research completion detection
- `gyoshu_snapshot` - Research state snapshots
- `notebook_search` - Search across notebooks
- + 9 more
๐ Full Changelog
- See [CHANGELOG.md](https://github.com/Yeachan-Heo/My-Jogyo/blob/main/CHANGELOG.md) for details.
v1.0.0 - Initial Release ๐v1.0.0
โจ โจ Features
- ๐ฌ Hypothesis-Driven Research โ Structure your work with `[OBJECTIVE]`, `[HYPOTHESIS]`, `[FINDING]` markers
- ๐ Persistent Python REPL โ Variables survive across sessions, just like a real Jupyter kernel
- ๐ Auto-Generated Notebooks โ Every experiment is captured as a reproducible `.ipynb`
- ๐ค Autonomous Mode โ Set a goal, walk away, come back to results
- ๐ AI-Powered Reports โ Turn messy outputs into polished research narratives
- ๐ Session Management โ Continue, replay, or branch your research anytime
๐ฆ ๐ Quick Install
- ```bash
- curl -fsSL https://raw.githubusercontent.com/Yeachan-Heo/My-Jogyo/main/install.sh | bash
- ```
๐ฆ ๐ Quick Start
- ```bash
- opencode # Start OpenCode
- /gyoshu # Meet the Professor
- /gyoshu analyze customer churn # Start research
- /gyoshu-auto build ML model # Autonomous mode
- ```
๐ฆ Commands
- `/gyoshu` - Unified research command (status, search, continue, report, etc.)
- `/gyoshu-auto <goal>` - Autonomous research with bounded execution
๐ฆ Tools
- | Tool | Purpose |
- |------|---------|
- | `research-manager` | Manage research projects and notebooks |
- | `notebook-search` | Search across notebooks |
- | `python-repl` | Persistent Python execution |
- | `session-manager` | Runtime session management |
- | `notebook-writer` | Jupyter notebook generation |
- | `gyoshu-completion` | Research completion signaling |
- + 2 more
๐ฆ ๐ Requirements
- OpenCode v0.1.0+
- Python 3.10+
- ---
- Made with ๐ for researchers who'd rather think than type
