t1yOS Skills

Expert skill package for AI coding assistants working with the t1yOS Serverless platform.

Overview

t1yOS Skills equips AI coding assistants (Claude Code, Cursor, Windsurf, etc.) with comprehensive knowledge of the t1yOS cloud operating system, enabling them to assist developers more effectively with cloud function development, database operations, SDK integration, and platform resource management.

Why This Skill Matters

t1yOS is a unique Serverless platform with significant differences from standard Node.js or mainstream cloud platforms:

• Cloud functions use the .jsc extension, not .js
• Entry point is a free function onRequest(ctx), not module.exports
• Module system uses require(), not import/export
• No npm ecosystem — only built-in platform modules are available
• Templates use Go {{ . }} syntax, not EJS/Pug
• Has its own security signing and encryption mechanisms

Without Skills, AI provides advice based on generic Node.js/Serverless experience that doesn't follow platform conventions. With Skills, AI generates code that's correct and deployment-ready.

Get the Skill

Region	Repository
Global	github.com/t1yOS/t1y-skills
China	gitee.com/t1yOS/t1y-skills

Installation

1. Download the ZIP package from GitHub (or Gitee for users in China) and extract it
2. Place the extracted t1y-skills directory in the appropriate location:

AI Tool	Install Path
Claude Code	~/.claude/skills/t1yOS/
Cursor	Project root .claude/skills/ or .cursor/skills/
Windsurf	Project root .claude/skills/
VS Code (Copilot)	Project root .claude/skills/

The AI assistant will auto-detect and load the skill — no additional configuration needed.

Coverage

Domain	Content
Cloud Functions	.jsc file conventions, onRequest entry point, built-in modules, routing rules, template rendering
Cloud Database	NoSQL CRUD, SQL-to-BSON, aggregation pipelines, pagination, query operators
Security	HMAC-SHA256 request signing, AES-256-GCM encryption, JWT, password hashing
MCP Server	Configuration, tool catalog, multi-AI tool adaptation
Client SDKs	JS/TS, Android, Swift, Flutter, C#, Go — six-platform integration guide
HTTP Constants	Complete status codes, methods, and MIME type reference

Benchmark Results

Code generation pass rate comparison with and without the skill:

Metric	With Skill	Without Skill
Pass Rate	100% ± 0%	93% ± 6%
Duration	40.5s ± 10.7s	43.8s ± 43.8s
Stability	Zero variance	High variance

With the skill, pass rate improved from 93% to 100%, with zero execution variance.

Project Structure

t1y-os-skills/
├── SKILL.md                    # Main skill file (loaded on trigger)
├── README.md                   # Chinese README
├── README.en.md                # This file (English)
├── evals/
│   └── evals.json              # Evaluation test cases
└── references/                 # Reference docs (loaded on demand)
    ├── cloud-function.md       # Cloud Function API reference
    ├── database.md             # Database API reference
    ├── crypto.md               # Crypto module API reference
    ├── mcp.md                  # MCP Server tools reference
    ├── sdks.md                 # Multi-platform SDK integration guide
    ├── security.md             # Security mechanisms deep-dive
    └── constants.md            # HTTP constants reference

Using with MCP Server

t1yOS provides two AI-assistance capabilities:

• Skills (this repository): Teaches AI the right way to write code
• MCP Server: Gives AI tools to directly operate the platform

Use both together for the best experience — Skills ensure correctness, MCP enables one-click deployment and management.

MCP Server URL: https://ai.t1y.net/mcp

License

This is an official t1yOS platform skill package, synchronized with platform documentation updates.