map-framework

Plan-then-build AI coding for Claude Code & Codex CLI — you approve the plan before the model writes a line of code. SPEC → PLAN → TEST → CODE → REVIEW → LEARN

What is map-framework?

Plan-then-build AI coding for Claude Code & Codex CLI — you approve the plan before the model writes a line of code. SPEC → PLAN → TEST → CODE → REVIEW → LEARN

Overview of map-framework

What it is

MAP (map-framework) is a plan-then-build AI coding framework for Claude Code and Codex CLI. It requires the user to approve a plan before the model writes any code. The workflow follows SPEC → PLAN → TEST → CODE → REVIEW → LEARN. It is open source under the MIT license and categorized as a Developer Tool.

Key features

  • Plan-then-build workflow: user approves the plan before the model writes code.
  • Eliminates common failure modes: silent architecture decisions, large diffs, tests written around implementation mistakes, and lost context between sessions.
  • Moves engineering judgment earlier: write behavior, split work into small contracts, verify each stage, review against the spec, and save lessons for the next run.
  • Supports both Claude Code and Codex CLI with provider-specific commands (e.g., /map-plan for Claude, $map-plan for Codex).
  • Automatic stable updates enabled by default; normal MAP skills check at most once per project every 24 hours and apply patch/minor releases automatically, asking before a major release.
  • Core commands include /map-auto, /map-plan, /map-efficient, /map-check, /map-review, /map-learn, /map-understand, /map-debug, and others.
  • A case study (DevOpsConf 2026) reports a production Kubernetes Project Operator estimated at 90 days delivered in 7 days using MAP.

Use cases

Good fits: complex backend features, Kubernetes controllers and operators, internal platform tooling, API/CRD/domain-model changes with invariants, and refactoring with a meaningful test harness. Poor fits: typos and tiny edits, small one-off scripts, product ideas where behavior is still unknown, broad rewrites without clear boundaries, and tasks cheaper to do directly than to plan.

Getting started

Install with:bashuv tool install mapify-cli# or with pippip install mapify-cliRequires Python 3.11+. Initialize your project:For Claude Code:bashcd your-projectmapify initclaudeFor Codex CLI:bashcd your-projectmapify init . --provider codexcodexThen run the golden path:Claude: /map-plan → /map-efficient → /map-check → /map-review → /map-learnCodex: $map-plan → $map-efficient → $map-check → $map-review → $map-learn

License and pricing

License: MIT. Pricing: open source.

Based on the project's public README and repository metadata.

Add this badge to your README

Featured on YunoDB Badge styles & help

Maintain this project? Show that it is featured on YunoDB.

Markdown[![Featured on YunoDB](https://yunodb.com/badge/azaliomap-framework.svg)](https://yunodb.com/tools/azaliomap-framework?utm_source=github&utm_medium=badge)
HTML<a href="https://yunodb.com/tools/azaliomap-framework?utm_source=github&utm_medium=badge"><img src="https://yunodb.com/badge/azaliomap-framework.svg" alt="Featured on YunoDB" height="20"></a>