lsp-client
v0.1.0Agent-Native Ecosystem

IntelliSense is designed for editors, leveraging the Language Server Protocol to provide advanced code capabilities; however, these features are difficult for LLMs to utilize directly.

Designed for
Coding Agent driven by LLMs (like Claude Code) execute programming tasks in repositories. This project enables these agents to enjoy the benefits of IntelliSense just like human programmers.

Transforming the Language Server Protocol (LSP) into actionable intelligence for AI Coding Agents.

openskills install lsp-client/lsp-skill --global
See https://github.com/numman-ali/openskills for OpenSkills documentation.
Get Started View on GitHub

Repository-Scale Intelligence

See how Coding Agents use LSP to understand your codebase in seconds.

Claude Code — LSAP Agent
* Welcome to Claude Code!

Installation

Pick your AI coding tool and follow the tool-specific steps.

Universal (Recommended)
Install using OpenSkills CLI
Steps
  • Install OpenSkills CLI via npm
  • Open your project directory
  • Install the LSP skill using openskills
Command
npm install -g openskills && openskills install lsp-client/lsp-skill
After running it, you may need to restart the tool to pick up the change.

System Architecture

From Protocol to Intelligence. A layered stack evolving raw capabilities into semantic understanding.

01
The Foundation
LSP Client
Industrial-strength Python SDK. Async-native architecture built on AnyIO for fine-grained protocol control.
  • Protocol Native
  • Async-First
  • Container Registry
02
The Abstraction
LSAP
Language Server Agent Protocol. Semantic abstraction layer transforming granular LSP ops into agent-ready snapshots.
  • Agent-Centric
  • Semantic Context
  • RPC -> Markdown
03
The Toolchain
LSP CLI
Universal command-line interface bringing LSP capabilities to any terminal pipeline or automation script.
  • Human Readable
  • Universal Interface
  • Scriptable
04
The Intelligence
LSP Skill
High-level repository analysis capabilities designed specifically for Coding Agents like Claude.
  • Deep Understanding
  • Cross-file Analysis
  • Plugin Ready

Extensibility

A three-tier extensibility design ensuring continuous growth and adaptation to new needs.

01
LSP Client
Foundation Expansion
The underlying library continuously expands support for more language ecosystems (Python, Rust, Go, etc.) and integrates new LSP 3.17 standard capabilities like Type Hierarchy and Call Hierarchy.
  • Broader Language Support
  • LSP 3.17 Specification Coverage
02
LSAP Protocol
Capability Expansion
Continuously designs composed capabilities like Relation API and Impact Analysis, while optimizing Markdown output formats based on Progressive Disclosure for better LLM reasoning.
  • Advanced Analysis Capabilities
  • Optimized Output Formats
03
Best Practice System
Scenario Expansion
A modular system enabling community contributions of domain-specific workflows (Frontend, Backend) and framework specializations (Django, React) to cover every development scenario.
  • Domain Expert Knowledge
  • Framework Specialization

Current Supported Languages

Advanced repository analysis powered by industry-standard Language Servers.

Python

Basedpyright

TypeScript

TypeScript Language Service

Go

Gopls

Rust

Rust Analyzer

Deno

Deno Language Server

More server support coming soon

We are continuously expanding our ecosystem with more languages and advanced LSP capabilities.