SkillsLLM
CategoriesBlogAI NewsAbout
HomeAI AgentsNexus-skills

Nexus-skills

by Haaaiawd

Pending

AI-native codebase intelligence skills — generate persistent .nexus-map/ knowledge bases and instantly query file structure, dependency graphs, and change impact. Built for Copilot, Cursor, and any tool-calling LLM. 面向 AI 编程助手的代码库感知技能——生成持久化知识图谱,精准查询文件结构、依赖关系与改动影响半径。

101stars
5forks
Python
Added 3/15/2026
View on GitHubDownload ZIP
AI Agentsagent-skillsagentic-aiai-codingai-toolsastclaude-codeclaude-skillsclinecode-analysiscodebase-intelligencecursordependency-graphgithub-copilotlegacy-coderefactoringskillsspectree-sittervibe-coding
Installation
# Add to your Claude Code skills
git clone https://github.com/Haaaiawd/Nexus-skills
README.md
<p align="center"> <img src="Icon.png" alt="nexus-skills" width="96" height="96"> </p> <h1 align="center">nexus-skills</h1> <p align="center"> 为 AI Agent 生成可持续复用的代码仓库知识库,并提供精准的代码结构查询能力。<br> Build a persistent knowledge map of any codebase. Query its structure with precision. Every AI session starts smarter. </p> <p align="center"> <a href="README.zh-CN.md">中文文档</a> </p>

Two Skills, One Philosophy

| Skill | What It Does | When to Use | |-------|-------------|-------------| | nexus-mapper | Analyzes the full repository and writes a persistent .nexus-map/ knowledge base for future AI sessions | When starting on an unfamiliar repo, onboarding a team, or preparing for major architectural work | | nexus-query | Queries file structure, reverse dependencies, change impact radius, and coupling hotspots directly from AST data | During active development — before interface changes, refactoring sprints, or legacy codebase navigation |

Both skills use the same underlying scripts. nexus-query can reuse a .nexus-map/ produced by nexus-mapper, or generate its own ast_nodes.json on demand.


nexus-mapper

nexus-mapper is a repository-mapping skill for AI agents. It analyzes a local codebase, writes a persistent .nexus-map/ knowledge base, and gives the next session a concrete place to start instead of forcing it to rediscover architecture from scratch.

This is not a generic "summarize the repo" prompt. The skill runs a gated PROBE workflow, challenges its own first-pass assumptions, and only then writes final assets. That design matters: it reduces the usual AI failure mode of turning first impressions into fake certainty.

.nexus-map/
├── INDEX.md              ← Load this first. Full architectural context, under 2000 tokens.
├── arch/
│   ├── systems.md        ← Every subsystem: what it owns, exactly where it sits in the repo.
│   ├── dependencies.md   ← How components connect. Rendered as a Mermaid dependency graph.
│   └── test_coverage.md  ← Static test surface: what is tested, what is not, and where evidence is thin.
├── concepts/
│   ├── concept_model.json ← Machine-readable knowledge graph. Structured for programmatic use.
│   └── domains.md        ← The domain language this codebase speaks, in plain terms.
├── hotspots/             ← Present when git metadata is available.
│   └── git_forensics.md  ← Files that change constantly, and pairs that always change together.
└── raw/                  ← Source data: AST nodes, git statistics, filtered file tree.

INDEX.md is the entry point and routing hub. After reading it, load all five companion files before taking action — they are intentionally kept short (typically under 5000 tokens combined).


nexus-query

nexus-query gives precise, instant answers to structural questions during active development — without reading the entire codebase.

# File skeleton: classes, methods, line numbers, imports
python skills/nexus-query/scripts/query...
Comments (0)
to leave a comment.

No comments yet. Be the first to share your thoughts!

Related Skills

gemini-cli

by google-gemini

An open-source AI agent that brings the power of Gemini directly into your terminal.
97,754
12,257
TypeScript
AI Agentsaiai-agents
View details
everything-claude-code

by affaan-m

The agent harness performance optimization system. Skills, instincts, memory, security, and research-first development for Claude Code, Codex, Opencode, Cursor and beyond.
76,097
9,506
JavaScript
AI Agentsai-agentsanthropic
View details
context7

by upstash

Context7 Platform -- Up-to-date code documentation for LLMs and AI code editors
49,063
2,314
TypeScript
MCP Serversllmmcp
View details
awesome-claude-skills

by ComposioHQ

A curated list of awesome Claude Skills, resources, and tools for customizing Claude AI workflows
44,141
4,446
Python
AI Agentsagent-skillsai-agents
View details
chatgpt-on-wechat

by zhayujie

CowAgent是基于大模型的超级AI助理,能主动思考和任务规划、访问操作系统和外部资源、创造和执行Skills、拥有长期记忆并不断成长。同时支持飞书、钉钉、企业微信应用、微信公众号、网页等接入,可选择OpenAI/Claude/Gemini/DeepSeek/ Qwen/GLM/Kimi/LinkAI,能处理文本、语音、图片和文件,可快速搭建个人AI助手和企业数字员工。
42,209
9,826
Python
AI Agentsaiai-agent
View details
ui-ux-pro-max-skill

by nextlevelbuilder

An AI SKILL that provide design intelligence for building professional UI/UX multiple platforms
41,523
4,026
Python
CLI Toolsai-skillsantigravity
View details