Start with AgentKit
Recommended for new setups. Learn the product model, then install the ak CLI.
Welcome to VividKit Guides
Choose AgentKit for new setups, keep ClaudeKit for existing ones, or migrate when you are ready.
Continue with ClaudeKit
Existing ClaudeKit setups remain supported. Keep using the guides and commands you already know.
Migrate from ClaudeKit to AgentKit
Back up configuration, review the differences, and migrate when ready. Migration is optional, not forced.
Tools & methods
Runtime tools and practical workflows. Some pages still assume ClaudeKit command syntax.
Commands & Workflows
Find the right command, browse the catalog, and use proven workflow recipes
AI Tools & Runtimes
Use ClaudeKit across providers, IDEs, remote control, and other AI coding tools
Visualization Tools
Visual tools for diagrams and browser review — architecture, workflows, and UI checkpoints.
Debug, Recover & Polish
Recover context, debug from logs, and improve UI output
Deals & Promotions
Limited timeSave on your AI stack
Hand-picked discounts on AI services and dev tools to cut your stack costs. Limited-time offers — grab them before they expire.
Inside ClaudeKit
Deeper topics for real workflows
This section will collect focused deep dives into specific ClaudeKit topics, so readers can understand them more clearly and apply them more effectively in their own workflow.
Getting Started with Claude Code & ClaudeKit Engineer Kit
Set up ClaudeKit, learn the 7 core commands, and apply best practices to ship features and fix bugs faster with your AI development partner.
Unlocking /ck:plan with --deep mode and --tdd flag
When to use --deep, when to add --tdd, and when a lighter approach is enough. Includes decision matrix and best practices.
Guard rails: the brake layer between model and action
When an agent can read files, run shells, edit code and ship PRs, one prompt line is not enough. Breaking down the 4 guard layers, 7 guard-rail groups, and the gaps that remain in ClaudeKit.