MDFlow Extension

VS Code companion for mdflow.dev — the CLI that turns markdown files into executable AI agents.

What is MDFlow?

MDFlow lets you write AI prompts as markdown files and run them as CLI commands:

review.claude.md                 # Run with Claude
commit.gemini.md "fix auth bug"  # Run with Gemini
git diff | explain.claude.md     # Pipe through any command

The filename determines the AI command, and YAML frontmatter becomes CLI flags:

# review.claude.md
---
model: opus
dangerously-skip-permissions: true
---
Review this code for bugs and suggest improvements.

@./src/**/*.ts

What This Extension Does

This extension provides intelligent autocomplete and documentation for mdflow frontmatter:

Autocomplete — Suggests valid YAML keys and values as you type
Inline Descriptions — See what each option does without leaving the dropdown
Hover Documentation — Full descriptions, types, defaults, and examples on hover
Built-in Schemas — Pre-configured support for all mdflow-compatible CLIs:
- *.claude.md — Claude Code CLI
- *.gemini.md — Gemini CLI
- *.copilot.md — GitHub Copilot CLI
- *.codex.md — Codex CLI

Installation

Install from the VS Code Marketplace or OpenVSX.

Or search for "MDFlow" in VS Code extensions.

Usage

Just create a file with a CLI suffix (e.g., task.claude.md) and start typing in the frontmatter:

---
model: |  ← autocomplete shows: sonnet, opus, haiku...
---

The extension automatically activates for any .md file matching the CLI patterns.

Custom Schemas

For non-CLI markdown files, you can map custom JSON schemas:

{
  "mdflow.schemas": {
    "~/schemas/blog.schema.json": "**/blog/**/*.md",
    "./schemas/docs.schema.json": "docs/**/*.md"
  }
}

License

MIT

MDFlow

johnlindquist