vt-agent-skills-standard

v1.9.2

Published

3 months ago

A CLI to manage and sync AI agent skills standard for Cursor, Claude, Copilot, Windsurf, and more.

0High
0Medium
0Low

vtt-vn

ai agent cursor claude copilot cli antigravity

Agent Skills Standard CLI 🚀

The engine behind High-Density AI coding. Command your AI assistants with professional standards.

The agent-skills-standard CLI is the official command-line tool to manage, sync, and version-control engineering standards across all major AI agents (Cursor, Claude Code, GitHub Copilot, Gemini, Roo Code, OpenCode, and more).

💡 What does this tool do?

If the Agent Skills Standard is the "instruction manual" for your AI, this CLI is the delivery truck that brings those instructions to your project.

Why you need this CLI

For Developers: No more copy-pasting .cursorrules or manual file management. One command keeps your AI updated.
For Non-IT/Tech Leads: Quickly set up a new project with the same professional standards used by senior engineers.
For Teams: Ensure every developer’s AI tool (Cursor, Claude, Copilot) behaves the same way across the entire codebase.

⚡ The Problem: "The Context Wall"

Modern AI coding agents are powerful, but they have major flaws:

Memory Drain: Giant rule files consume 30% - 50% of the AI's memory, making it less effective for actual coding.
Version Chaos: Team members often have different "best practices," leading to inconsistent code.
Wordy Prose: Human-style instructions are token-heavy and often ignored by AI during complex logical tasks.

Agent Skills Standard solves this by treating prompt instructions as versioned dependencies, similar to how you manage software libraries.

🛠 The Solution: Digital DNA for AI

Agent Skills Standard treats instructions as versioned dependencies, much like software libraries.

🎯 Smart Loading: We use a "Search-on-Demand" pattern. The AI only looks at detailed examples when it specifically needs them, saving its memory for your actual code.
🚀 High-Density Language: We use a specialized "Compressed Syntax" that is 40% more efficient than normal English. This means the AI understands more while using fewer resources.
🔁 One-Click Sync: A single command ensures your AI tool stays up-to-date with your team's latest standards.

🛡️ Security & Trust

We take security seriously. Here is what you need to know about how the CLI works:

No Hidden Scripts: The CLI sync command only downloads text files (Markdown/JSON). It does not download or execute binaries, scripts, or unknown code.
No "Code Injection": When we say "injection", we mean Prompt Injection (adding context to the AI's conversation history), NOT code injection (running malware).
Transparent Operations:
- The CLI fetches a standard directory structure from the official registry.
- It copies these text files to your local .cursor/skills or .agent/skills folder.
- It updates AGENTS.md (a text file index).
- That is it. No background services, no daemons, no hidden network calls.

📊 Efficiency & Benchmark

Every skill delivered by this CLI is audited for its footprint in the AI's context window.

| Metric | Savings (vs. Heavy Prompt) | Avg. Footprint | Quality | | ------------------ | -------------------------: | -------------- | ------- | | Global Average | 88% Savings | ~457 tokens | 8.9/10 |

[!IMPORTANT] Context is Currency: By reducing instruction overhead by 88%, you free up your AI's memory and budget for complex logic and large codebases.

🛡️ Multi-Agent Support: Out-of-the-box mapping for Cursor, Claude Dev, GitHub Copilot, and more.
📦 Modular Registry: Don't load everything. Only enable the skills your project actually uses.
⚡ Proactive Activation (Universal): Generates a compressed index in AGENTS.md for 100% activation reliability across Cursor, Windsurf, Claude Code, and more.
🔄 Dynamic Re-detection: Automatically re-enables skills if matching dependencies are added.
🔒 Secure Overrides: Lock specific files so they never get overwritten.
📊 Semantic Tagging: Skills tagged with triggers for exact application.
🤖 Agent Workflows: Sync executable workflows (.md files) that agents can follow to perform complex tasks.

📜 Benchmark History

| Version | Date | Skills | Avg Tokens | Savings (%) | Report | | ------- | ---------- | ------ | ---------- | ----------- | -------------------------------------- | | v1.9.2 | 2026-03-07 | 228 | 458 | 87% | Report | | v1.9.1 | 2026-03-07 | 228 | 458 | 87% | Report | | v1.9.0 | 2026-03-05 | 228 | 457 | 88% | Report | | v1.8.0 | 2026-03-02 | 228 | 443 | 88% | Report |

� Token Economy & Optimization

To ensure AI efficiency, this project follows a strict Token Economy. Every skill is audited for its footprint in the AI's context window.

📏 Our Standards

High-Density: Core rules in SKILL.md are kept under 100 lines.
Efficiency: Target < 500 tokens per primary skill file.
Progressive Disclosure: Heavy examples, checklists, and implementation guides are moved to the references/ directory and are only loaded by the agent when specific context matches.

🔒 Privacy & Feedback

Feedback Reporter Skill

By default, the CLI syncs a common/feedback-reporter skill that enables you to report when AI makes mistakes or when skill guidance needs improvement. This helps us improve skills for everyone.

What Gets Shared (Only if AI Agent Report):

Skill category/name
Issue description (written by you or generated by AI)
Skill Instruction: Exact quote from the skill that was violated
Actual Action: What the AI did instead
Decision Reason: Why the AI chose that approach
Optional context (framework version, scenario)
Optional AI Model name
NO code, NO project details, NO personal information

How to Opt-Out: Add to .skillsrc:

skills:
  common:
    exclude: ['feedback-reporter']

How to Report Issues: Use structured comments:

// @agent-skills-feedback
// Skill: react/hooks
// Issue: AI suggested unsafe pattern
// Suggestion: Add guidance for this case

Privacy First: We never collect usage telemetry or analytics. Feedback is only shared if you explicitly trigger it.

Manual Feedback

If you notice a skill needs improvement, you can manually send feedback using:

ags feedback

Or via structured comments in your code:

// @agent-skills-feedback
// Skill: react/hooks
// Issue: AI suggested unsafe pattern
// Suggestion: Add guidance for this case

🚀 Installation

You can run the tool instantly without installing, or install it globally for convenience:

# Use instantly (Recommended)
npx agent-skills-standard@latest sync

# Or install globally
npm install -g agent-skills-standard

# Use the short alias
ags sync

🛠 Basic Commands

1. Setup Your Project

Run this once to detect your project type and choose which "skills" you want your AI to have.

npx agent-skills-standard@latest init

2. Boost Your AI

Run this to fetch the latest high-density instructions and install them into your hidden agent folders (like .cursor/skills/ or .github/skills/).

npx agent-skills-standard@latest sync

⚙️ Configuration (`.skillsrc`)

The .skillsrc file allows you to customize how skills are synced to your project.

registry: https://github.com/HoangNguyen0403/agent-skills-standard
agents: [cursor, copilot]
skills:
  flutter:
    ref: flutter-v1.1.0
    # 🚫 Exclude specific sub-skills from being synced
    exclude: ['getx-navigation']
    # ➕ Include specific skills (supports cross-category 'category/skill' or 'category/*' syntax)
    include:
      - 'bloc-state-management'
      - 'react/hooks'
      - 'common/*'
    # 🔒 Protect local modifications from being overwritten
    custom_overrides: ['bloc-state-management']
  # 🤖 Optional: Sync workflows to .agent/workflows/
  workflows: true

Key Options

exclude: A list of skill IDs to skip during synchronization.
include: A list of skill IDs to fetch. Supports:
- Relative Path: bloc-state-management (from current category)
- Absolute Path: react/hooks (pull specific skill from another category)
- Glob Path: common/* (pull ALL skills from another category)
custom_overrides: A list of skill IDs that the CLI should never overwrite.
ref: Specify a specific version or tag for the skills.

✨ Key Features

🎯 Efficiency First: Uses a "Search-on-Demand" pattern that only loads information when the AI needs it, saving its "brain power" for your code.
🚀 High-Density Instructions: Optimized syntax that is 40% more compact than standard English.
🛡️ Universal Support: Works out-of-the-box with Cursor, Claude, GitHub Copilot, and more.
🔒 Secure Protection: Mark specific files as "Locked" (overrides) so the CLI never changes your custom tweaks.
🧪 Production-Grade Reliability: Guarded by a 100% statement coverage test suite and strict CI enforcement.

🔹 Supported Stacks (Token Audited)

| Category | Typical Savings | Avg. Footprint | Status | | :-------------- | :-------------- | :------------------------------------------------------------------------------- | :------------ | | Flutter | 88% | 422 | ✅ Stable | | NestJS | 85% | 518 | ✅ Production | | Next.js | 89% | 431 | ✅ Stable | | React | 90% | 401 | ✅ Stable | | Golang | 90% | 375 | ✅ Stable | | Android | 92% | 295 | ✅ Production | | iOS | 90% | 411 | ✅ Production | | Laravel | 91% | 361 | ✅ Stable | | Spring Boot | 92% | 369 | ✅ Stable |