cloud-init-crafter（云初始化工匠）

Cloud-init Crafter

This skill provides a structured framework for generating and maintaining cloud-init configurations. It utilizes a library of baseline templates, localized reference guides, and automated validation scripts to ensure production-ready deployments.

Core Principles

1. Plan Before Execution: Always define a concise implementation plan (listing target modules, target format, baseline templates, and validation logic) before generating code.
2. Format Awareness: Recognize that cloud-init accepts multiple formats. Default to #cloud-config (YAML 1.1) but proactively suggest shell scripts, boothooks, or MIME archives when the user's operational timing requires it.
3. Security by Default: Prohibit plaintext passwords. Use ssh_authorized_keys or hashed_passwd exclusively. Ensure sensitive instance-data is accessed securely.
4. Idempotency: Ensure all runcmd, bootcmd, and custom scripts are idempotent to allow safe re-execution.

Local References (Load On-Demand)

Load these reference documents on-demand when specific technical details, formats, or troubleshooting steps are required to save token context.

Format Guides

Load these when defining the structure of the user-data payload or combining multiple execution types:

• Cloud Config Format - Standard YAML configurations.
• Cloud Boothook Format - Early-boot shell scripts (Network stage).
• User-data Script Format - Late-boot standard shell scripts.
• Include File Format - Referencing external URLs.
• Jinja Template Format - Header requirements for Jinja rendering.
• Gzip Compressed Content - Handling binary compressed payloads.
• Cloud Config Archive - YAML-based multi-part alternatives.
• MIME Multi-part Archive - Combining multiple formats into a single payload.

Data & Templating

• Instance Data - Standardized v1 keys, JSON structure, and Jinja templating usage for dynamic, environment-aware configurations.

Core Concepts & Guidelines

• Concepts and Boot Stages - Overview of cloud-init principles and the five boot stages (Detect, Local, Network, Config, Final).
• Cloud-config API - Detailed syntax, configuration schema, and examples for common configuration modules.
• Best Practices - Strategic guidelines for writing robust, idempotent, and maintainable cloud-init scripts.
• Security Guidelines - Critical security hardening, credential management, and SSH host key verification.
• CLI Reference - Comprehensive command-line interface guide for cloud-init and cloud-init-per.
• Error Cheatsheet - Troubleshooting workflows, log paths, status codes, and common failure modes.

Platform Launching Guides

• QEMU Launching - Guide for local instance testing using QEMU and the NoCloud datasource.
• LXD Launching - Guide for testing with LXD containers.
• Multipass Launching - Guide for launching VMs on Linux, Windows, or macOS.
• Libvirt Launching - Automating seed image creation and deployment.
• WSL Launching - Requirements for Ubuntu 24.04+ on Windows Subsystem for Linux.

Assets (Baseline Templates)

Use these templates as a starting point for new configurations:

• Minimal Template - Essential SSH-only access with password authentication disabled.
• Default Template - Standard baseline for production including users, SSH keys, and package updates.
• Complete Template - Comprehensive setup including disk partitioning, file writing, and complex orchestration.

Validation Tools

• Config Validator - A Python script (PEP 723) to offline validate YAML syntax and mandatory headers.

Official Documentation

Refer to these external resources for the most granular specifications:

• Official Documentation
• Cloud-config Modules Reference
• Datasources Reference

Standard Operating Procedure (SOP)

1. Discovery & Planning: Analyze the user's requirements. Determine the correct Boot Stage needed (early network vs. final execution) and select the appropriate format (e.g., #cloud-config, #cloud-boothook, or MIME). Decide if dynamic instance-data via Jinja is required. Output a concise plan.
2. Context Loading: Silently load the relevant format guide from references/formats/ and references/instance-data.md if templating is needed.
3. Execution: Generate or modify the configuration block. Ensure strict adherence to the loaded format's specific headers, syntax, and security constraints.
4. Validation: Recommend validation via cloud-init schema -c <file> --annotate or the provided scripts/validate_config.py tool.