improve — What is it?

The shadcn/improve project is an open-source tool designed to audit codebases and generate detailed implementation plans for efficient execution by cheaper models.

⭐ 7,911 Stars 🍴 335 Forks MIT Author: shadcn
Source: README View on GitHub →

Why it matters

The project is gaining attention due to its innovative approach of using a capable model for codebase understanding and planning, while delegating execution to cheaper models. This reduces costs and increases efficiency. The project stands out for its focus on self-contained plans, verification gates, and hard boundaries, making it unique in its execution planning capabilities.

Source: README, How it works section

Core Features

Codebase Audit

Audits the codebase across multiple categories such as correctness, security, performance, and tech debt, providing detailed findings with evidence and impact assessments.

Source: README, How it works section
Plan Generation

Generates executable plans for code changes, ensuring all context is inlined and verification gates are set for each step, making the plans self-contained and machine-checkable.

Source: README, What makes the plans executable section
Execution and Reconciliation

Supports execution of plans by cheaper models and reconciliation of the execution results, allowing for iterative improvements and updates to the plans.

Source: README, Closing the loop section

Architecture

The architecture of shadcn/improve is modular, with distinct components for mapping the codebase, auditing, planning, execution, and reconciliation. It uses parallel subagents for auditing and employs a design that ensures plans are self-contained and executable by weaker models.

Source: README, How it works section

Project Knowledge Graph

Knowledge graph: project (center) + core features (inner hexagons) + key dependencies (outer chips) Not enough information.Not enough inf… Codebase Audit Plan Generation Execution and ReconciliationExecution and Recon… improve Project Core feature Key dependency

Center: project; inner ring: core feature modules; outer ring: key dependencies. Auto-generated from core_features and tech_stack.key_deps.

Tech Stack

LanguageNot enough information.FrameworkNot enough information.
Not enough information.
Not enough information.
Source: Code tree, Dependency files

Quick Start

npx skills add shadcn/improve Works in any agent that supports [Agent Skills](https://agentskills.io) format. The plans it writes are plain markdown, so any agent (or human) can pick them up.
Source: README Installation/Quick Start

Use Cases

The project is suitable for developers and technical teams working on large codebases who need to efficiently manage code changes, improve code quality, and reduce costs associated with execution.

Source: README

Strengths & Limitations

Strengths

  • Strength 1: Reduces costs by using cheaper models for execution.
  • Strength 2: Improves code quality through comprehensive audits and detailed plans.
  • Strength 3: Facilitates efficient management of code changes and updates.

Limitations

  • Limitation 1: Requires an agent that supports the Agent Skills format.
  • Limitation 2: The project is still in development and may have limitations in its current state.
Source: README, How it works section

Latest Release

Not enough information.

Source: GitHub Releases

Verdict

shadcn/improve is a promising project for teams looking to optimize codebase management and execution. Its innovative approach to using a capable model for planning and cheaper models for execution offers a unique solution for reducing costs and improving efficiency, although it may require some adaptation to existing workflows and tools.

Frequently Asked Questions

What is improve?

The shadcn/improve project is an open-source tool designed to audit codebases and generate detailed implementation plans for efficient execution by cheaper models.

What are the main features of improve?

improve's core features include: Codebase Audit, Plan Generation, Execution and Reconciliation.

Why is improve trending?

The project is gaining attention due to its innovative approach of using a capable model for codebase understanding and planning, while delegating execution to cheaper models.

What is improve used for?

The project is suitable for developers and technical teams working on large codebases who need to efficiently manage code changes, improve code quality, and reduce costs associated with execution.

Transparency Notice
This page is auto-generated by AI (a large language model) from the following public materials: GitHub README, code tree, dependency files and release notes. Analyzed at: 2026-07-07 18:34. Quality score: 85/100.

Data sources: README, GitHub API, dependency files