开源雷达 约 17 分钟

开源雷达 · 2026-08-23

Collected 2026-08-23 09:45 · 9 projects

Data: GitHub Trending and GitHub REST metadata. Method: deterministic ranking followed by DeepSeek structured analysis.

快速浏览

anthropics/claude-code

Claude Code: An agentic coding tool that lives in your terminal

中等

Language Python · License 未声明 · Stars 142548 · Forks 22838 · Recent +127 stars today

Problem: Developers spend significant time on routine coding tasks, understanding complex codebases, and managing git workflows, which reduces overall productivity.

Approach: Claude Code is an agentic coding tool that operates in the terminal, understands the codebase, and executes routine tasks, explains complex code, and handles git workflows through natural language commands. It can be used in the terminal, IDE, or by tagging @claude on GitHub.

Audience: Developers who want to speed up coding tasks and manage projects via natural language in the terminal.

GitHub Trending rank #8; 127 stars today.

Shubhamsaboo/awesome-llm-apps

100+ Open-Source AI Agents, Skills, and RAG Apps

中等

Language Python · License Apache License 2.0 · Stars 133608 · Forks 19650 · Recent no verified recent signal

Problem: Developers need ready-to-use, production-style AI agents and RAG applications without building from scratch.

Approach: A curated collection of 100+ hand-built, tested AI agents, agent skills, and RAG apps, covering starter, advanced, always-on, and multi-agent team scenarios. Includes step-by-step tutorials and quick-start commands.

Audience: Developers and AI enthusiasts looking for practical, open-source AI agent templates and examples.

No recent GitHub Trending rank or daily star signal is available.

anthropics/skills

Anthropic's Agent Skills: Teach Claude Repeatable Tasks

中等

Language Python · License 未声明 · Stars 171038 · Forks 20324 · Recent no verified recent signal

Problem: Claude needs to perform specialized tasks in a repeatable way, such as creating documents with brand guidelines or analyzing data with specific workflows.

Approach: Skills are folders containing instructions, scripts, and resources that Claude loads dynamically. The repository provides examples, a specification, and a template for creating custom skills.

Audience: Developers and enterprises using Claude to automate tasks, and those interested in building custom skills.

No recent GitHub Trending rank or daily star signal is available.

nilbuild/developer-roadmap

Interactive Roadmaps for Developer Career Growth

容易

Language TypeScript · License Other · Stars 365179 · Forks 44834 · Recent no verified recent signal

Problem: Developers often struggle to navigate the vast and ever-changing landscape of technologies and career paths, lacking clear guidance on what to learn and in what order.

Approach: Provides community-driven, interactive roadmaps, best practices, and questions covering a wide range of topics from frontend, backend, DevOps, AI, and more, with clickable nodes that link to detailed content.

Audience: Developers at all levels, from beginners to experienced professionals, seeking structured learning paths and career guidance.

No recent GitHub Trending rank or daily star signal is available.

deepseek-ai/deepseek-harness

DeepSeek Harness: An open-source agent harness where everything is a plugin

中等

Language TypeScript · License MIT License · Stars 185175 · Forks 20487 · Recent no verified recent signal

Problem: Building AI agents often requires rigid, monolithic frameworks that are hard to extend or customize.

Approach: DeepSeek Harness (dsh) is an open-source agent harness developed by DeepSeek AI. It uses a plugin-based architecture powered by Cordis, where every component is a plugin, enabling flexible composition and spatiotemporal composability.

Audience: Developers and researchers building AI agents, especially those interested in plugin-based architectures and the Cordis ecosystem.

No recent GitHub Trending rank or daily star signal is available.

mattpocock/skills

Matt Pocock's Agent Skills for Real Engineering

中等

Language Shell · License MIT License · Stars 232116 · Forks 19812 · Recent no verified recent signal

Problem: AI coding agents often fail to align with user intent, are overly verbose, produce code that doesn't work, and accelerate software entropy, leading to complex and hard-to-change codebases.

Approach: A collection of small, composable, and adaptable skills that work with any model, focusing on engineering fundamentals like grilling sessions, shared language, TDD, and architecture improvement.

Audience: Software engineers using AI coding agents like Claude Code, Codex, and others who want to improve code quality and maintainability.

No recent GitHub Trending rank or daily star signal is available.