claude-code-skill

Document Claude Code CLI architecture and development conventions for TypeScript and Bun.

1|Updated Apr 8, 2026
One-click install
npx skills add https://github.com/Tashima-Tarsh/Disha --skill claude-code-skill-tashima-tarsh
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: claude-code-skill
Source: https://github.com/Tashima-Tarsh/Disha/tree/main/disha/docs
Command: npx skills add https://github.com/Tashima-Tarsh/Disha --skill claude-code-skill-tashima-tarsh

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Contributing to the Claude Code CLI repository requires navigating a complex TypeScript codebase with strict architectural patterns, extensive tooling systems, and multi-layered state management without a centralized reference.

Core Features & Use Cases

  • Architecture Overview: Maps the 7-layer structure from CLI core to AI platform, including the 7-stage cognitive loop, service registry, and knowledge graph layers.
  • Pattern Library: Documents conventions for tool definitions, command structures, React/Ink components, permission systems, and feature flags.
  • Use Case: A developer adding a new agent tool can reference the buildTool pattern, permission check requirements, and UI rendering conventions to ensure consistency with the existing 40+ tools.

Quick Start

Use the claude-code-skill guide to review the QueryEngine architecture and implement a new tool by following the buildTool pattern with proper permission checks and UI renderers.

Frequently Asked Questions about claude-code-skill

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I add a new tool to the Claude Code CLI codebase?

To add a new tool to the Claude Code CLI, follow the buildTool pattern documented in the architecture guide to ensure proper permission checks, validation, and UI renderers consistent with the existing 40-plus tools.

What architecture does the Claude Code CLI use for its cognitive loop?

The Claude Code CLI architecture uses a 7-stage cognitive loop within a 7-layer structure, mapping the flow from the CLI core to the AI platform, including a service registry and knowledge graph layers.

Can I use React and Ink to build terminal UI components for Claude Code?

Yes, the Claude Code CLI uses React and Ink for its terminal UI. The pattern library documents conventions for building components that adhere to the project's state management and rendering standards.

Does the Claude Code CLI support multi-agent coordination and MCP protocol integration?

Yes, the Claude Code CLI supports multi-agent coordination patterns and MCP protocol integration. The architecture guide covers how these systems interact within the TypeScript and Bun codebase.

What conventions exist for defining permissions and feature flags in Claude Code?

The Claude Code CLI architecture guide documents specific conventions for permission systems and feature flagging, ensuring new commands and tools maintain strict validation and state management standards.

Why do I need an architecture guide for contributing to the Claude Code CLI repository?

Contributing to the Claude Code CLI repository requires navigating a complex TypeScript codebase with strict architectural patterns and multi-layered state management without a centralized reference, which this guide provides.