milknado

name	milknado
description	Mikado-method execution with milknado: dependency graphs, parallel ralph loops in worktrees, and code-review-graph context. Use when planning goals, running autonomous loops, or inspecting graph state.

Milknado

When to use

• Decomposing a goal into ordered work (milknado plan)
• Running ready leaves as parallel agent loops (milknado run)
• Inspecting node status or adding prerequisites (milknado status, milknado add-node)

Configuration

• milknado.toml: agent_family (claude | cursor | gemini | codex), plus optional planning_agent / execution_agent overrides.
• Built-in family defaults use stdin for planning prompts and ralphify-compatible print-mode binaries for execution.

MCP

This plugin ships mcp.json registering the milknado-mcp stdio server (milknado_graph_summary, milknado_todo_add). Install the milknado package in the workspace (e.g. path dep in pyproject.toml) so uv run milknado-mcp resolves.