Turn any codebase, with its docs, SQL schemas, configs, and PDFs, into a queryable knowledge graph. A /graphify skill for Claude Code, Cursor, Codex, and Gemini CLI: local…
Skill Manager
Zero-dependency CLI to scan, recommend, deduplicate, audit, and visualize Claude Code / Codex skills and MCP servers.
Install
npm i -g aide-skill-managerInstall command taken from the project's README or registry entry. Always review code before running it.
README excerpt
Read full README on GitHub →skill-manager (skm)
English | 简体中文
Local skill / MCP governance for people who use multiple AI coding tools.
As your local skill collection grows,
skmanswers the questions that become hard to answer by hand: what is installed, which skill fits a task, what is duplicated or idle, where did it come from, and whether it is safe to update.
Who It Is For
skm is for AI developers who use Claude Code, Codex, Cursor, Gemini, WorkBuddy, Kimi, or more than one of them and have accumulated a growing local skill / MCP collection.
It is also useful for maintainers and teams that need a reproducible local baseline: source records, instance-scoped lock files, policy checks, and CI-friendly drift verification.
The product is local governance, not a remote skill marketplace or centralized team console. It is read-only by default, never updates skills automatically, and never executes a skill or MCP server.
The Problems It Solves
| The problem | What skm does |
|---|---|
| I do not know what is installed anymore | Scans skills and MCP servers across supported AI tools and builds one catalog |
| I know the task, but not the right skill | Recommends a skill with reasons and alternatives from local metadata |
| Duplicates, idle skills, and context cost keep growing | Finds duplicate entities, stale usage, and expensive MCP schemas, then suggests downshifts |
| A skill has no trustworthy source | Stores source metadata; missing sources can be entered manually or found through authorized GitHub discovery |
| An update might overwrite local work | Checks versions, commits, and package hashes; shows diffs and dry-run plans before writes |
| The local setup drifts over time | Creates instance-level lock files and verifies drift and policy baselines |
30-Second Start
npm i -g aide-skill-manager
# 1. See what is installed
skm scan
# 2. Describe a task when you do not know which skill to use
skm ask "convert a web page to Markdown"
# 3. Check recorded upstream sources for newer versions
skm outdated --online
Useful next commands:
skm dupes # find duplicates
skm audit # inspect real usage and static safety signals
skm sources missing # find missing upstream metadata
skm web # open the local Web dashboard
Optional bridge setup for Claude Code and Codex:
skm setup
skm setup installs the bundled skill-navigator bridge skill as an explicit write operation. It is not installed automatically with the CLI.
Source install for local development:
git clone https://github.com/GrubbyLee/skill-manager.git
cd skill-manager
node scripts/install.mjs
Four Common Workflows
1. Inventory: what is actually installed?
skm scan
skm
skm list
skm list --mcp
scan rebuilds the local catalog. The bare skm command groups findings by inventory, risks, usage, versions, lifecycle, duplicates, graph, and recommendations. Scanning does not read MCP env values or execute skills.
2. Recommendation: which skill fits this task?
skm ask "create image cards for Xiaohongshu"
skm recommend "markdown to html" --why
Similar plugs
Persistent Context Across Sessions for Every Agent – Captures everything your agent does during sessions, compresses it with AI, and injects relevant context back into future…
Graphs that teach > graphs that impress. Turn any code into an interactive knowledge graph you can explore, search, and ask questions about. Works with Claude Code, Codex,…
From vibe coding to agentic engineering - practice makes claude perfect
【低代码v2.0,一句话即可生成整个系统】企业级AI低代码平台,一键生成前后端代码甚至整个系统。 AI Skills 一句话画流程、设计表单、生成报表、大屏。内置 AI应用平台涵盖:AI聊天、知识库、流程编排、MCP插件等,兼容主流大模型。引领AI低代码「Skills 生成 → 在线配置 → 代码生成 → 手工合并->AI修改」开发模式,解决…
Open-source personal AI assistant & Agent Harness. Plans tasks, runs tools and skills, self-evolves with memory and knowledge. Multi-agent, multi-model, multi-channel.…