Installation¶

neuroflow is a plugin for Claude Code and GitHub Copilot CLI. Install it with whichever AI assistant you use.

Claude Code¶

From the marketplace¶

The easiest way to install neuroflow in Claude Code:

claude plugin marketplace add stanislavjiricek/neuroflow
claude plugin install neuroflow@neuroflow

Or from within an interactive Claude Code session:

/plugin marketplace add stanislavjiricek/neuroflow
/plugin install neuroflow@neuroflow

Local development (Claude Code)¶

Clone the repo and point Claude Code at it:

git clone https://github.com/stanislavjiricek/neuroflow
claude --plugin-dir ./neuroflow

This is the recommended way to contribute or test changes before they are released.

GitHub Copilot CLI¶

neuroflow installs natively into the GitHub Copilot CLI — no workarounds needed.

Install directly from GitHub¶

copilot plugin install stanislavjiricek/neuroflow

Or from within an interactive Copilot session:

/plugin install stanislavjiricek/neuroflow

Install from a registered marketplace¶

If stanislavjiricek/neuroflow is available in a registered marketplace (e.g. awesome-copilot):

copilot plugin install neuroflow@awesome-copilot

Uninstall (Copilot)¶

copilot plugin uninstall neuroflow

Verify installation¶

Once installed (in either tool), open any project folder and run:

/neuroflow:neuroflow

You should see neuroflow scan your project and offer to set up .neuroflow/ project memory.

MCP server requirements¶

neuroflow uses four MCP (Model Context Protocol) servers that are launched automatically via npx. No manual installation is needed — they are pulled from npm on first use.

Server	npm package	Purpose
PubMed / bioRxiv	`paper-search-mcp-nodejs`	Literature and preprint search
Miro	`@k-jarzyna/mcp-miro`	Visual collaboration boards
Context7	`@upstash/context7-mcp`	Library documentation lookup

One of these requires credentials — see the Integrations page for how to configure them.

System requirements¶

Requirement	Notes
Claude Code or GitHub Copilot CLI	Any recent version
Node.js	Required for MCP servers via `npx`
Python	Optional — only needed if you run preprocessing / analysis scripts the AI generates

Uninstall (Claude Code)¶

claude plugin uninstall neuroflow

Using on other platforms¶

neuroflow's skills and commands can also be adapted for other AI coding assistants.

GitHub Copilot (VS Code extension)¶

The VS Code Copilot extension does not share the same plugin system as the Copilot CLI, but you can make neuroflow's instructions available to Copilot agent mode via instruction files.

Option 1 — Project-level instructions (recommended):

Create a .github/copilot-instructions.md file in your project root and paste the contents of your relevant neuroflow skill files into it. Copilot agent mode reads this file automatically when working in the project.

<!-- .github/copilot-instructions.md -->
<!-- Paste relevant neuroflow skill content here -->

Option 2 — VS Code user instructions:

Open VS Code Settings → search for github.copilot.chat.codeGeneration.instructions → add a new instruction entry pointing to any neuroflow SKILL.md file (absolute path).

Option 3 — .instructions.md files in .github/instructions/:

VS Code Copilot also reads *.instructions.md files from .github/instructions/. Copy individual skill files there and rename them with the .instructions.md extension.

Limitations on VS Code Copilot extension: - No native slash command routing — commands must be triggered by natural language - No automatic .neuroflow/ project memory creation — do this manually or prompt Copilot Agent to create it - MCP server connections work the same way (configured in VS Code MCP settings)

Cursor¶

Cursor supports custom AI rules via the .cursor/rules/ directory (formerly .cursorrules). You can expose neuroflow skills to Cursor's AI assistant.

Setup:

Create a .cursor/rules/ directory in your project root
Copy the neuroflow skill files you want Cursor to follow into .cursor/rules/, renaming each to *.mdc:

# Example: add the neuroflow-core and review-neuro skills
cp path/to/neuroflow/skills/neuroflow-core/SKILL.md .cursor/rules/neuroflow-core.mdc
cp path/to/neuroflow/skills/review-neuro/SKILL.md .cursor/rules/review-neuro.mdc

Cursor will include these rules when the AI generates responses in agent mode.

Triggering commands in Cursor:

Cursor does not have a /command routing system. Instead, refer to commands by natural language: - "Follow the neuroflow flowie skill and set up a flowie for this project" - "Use the review-neuro skill to review this manuscript"

Project memory in Cursor:

Ask Cursor's agent to run the neuroflow interview manually: - "Read the neuroflow-core skill rules and set up .neuroflow/ project memory for this project"

Limitations on Cursor: - No automatic plugin installation — copy skill files manually - No marketplace integration - Command routing relies entirely on natural language + rules files