Commands Reference

Complete reference for all ContextKit CLI commands. You can use the short alias ck instead of contextkit.

Install & Setup

contextkit install

Install ContextKit in your current project. Creates .contextkit/ directory and configures integrations.

contextkit install

💡 Platform-Specific Installation

If .contextkit already exists (e.g., installed by another team member), add your specific AI tool integration:

contextkit cursorcontextkit vscodecontextkit claudecontextkit geminicontextkit codexcontextkit aidercontextkit continue

These commands only add the integration files for your specific AI tool without affecting the existing .contextkit/ directory.

contextkit status

Check installation status and verify which context files are loaded for AI prompts.

contextkit status

Shows which .md files will be loaded when you run ck "prompt"

contextkit analyze

Analyze your project and generate customized standards based on your actual codebase. Monorepo-aware: Automatically detects and prompts for frontend/backend scope.

contextkit analyze

Options:

  • --scope frontend|backend|both - Analyze specific scope (non-interactive)
  • --package <path> - Analyze specific package (e.g., apps/admin)
  • --non-interactive - Skip prompts

What it does: AI scans your codebase and generates content for skeleton standards files. ⚠️ You must review and edit the generated content manually.

📦 Monorepo Support:

For monorepos, automatically detects packages and prompts to analyze frontend, backend, or both. Generates separate standards when analyzing both.

Validation & Compliance

contextkit check

Validate installation and check policy compliance based on .contextkit/policies/policy.yml.

contextkit check

Options:

  • --strict - Treat warnings as errors
  • --verbose - Show detailed information

Corrections Logging

contextkit note

Add entries to the corrections log for tracking AI performance issues and improvements.

contextkit note "AI didn't follow testing standards"

Options:

  • --category <category> - Category (AI Behavior, Preferences, etc.)
  • --priority <priority> - Priority (HIGH, MEDIUM, LOW)
  • --task <task> - Related task description

Example: ck note "Good behavior observed" --category "Preferences" --priority LOW

Workflow Orchestration

contextkit run <workflow>

Execute structured workflows defined in .contextkit/instructions/core/.

contextkit run create-component

Options:

  • --interactive - Pause between steps for review

Workflows use XML-like tags (<process_flow>, <step>) for structured execution.

Registry & Versioning

contextkit publish

Package and publish your ContextKit configuration to the local registry for sharing across teams.

contextkit publish --name @company/react-standards --version 1.0.0

Creates a versioned archive in ~/.contextkit-registry.

contextkit pull <package>

Pull a ContextKit configuration from the registry.

contextkit pull @company/react-standards@1.0.0

Options:

  • --force - Overwrite existing files
  • --backup - Backup existing files before overwriting

Use @latest to pull the latest version.

Observability

contextkit dashboard

Start a web-based observability dashboard to visualize standards freshness, corrections log analytics, and policy compliance.

contextkit dashboard

Options:

  • --port <port> - Custom port (default: 3001)
  • --no-server - Display metrics in CLI only

Dashboard shows: standards freshness, corrections log statistics, policy compliance, and product context status.

Updates & Maintenance

contextkit update

Update ContextKit to the latest version.

contextkit update

AI Commands

contextkit <prompt>

Chat with AI using ContextKit context. Just type your prompt directly!

contextkit "create a button component"

✨ Shortcut: You can skip the "ai" keyword and just use ck "your prompt"

💡 How it works:

  • Automatically loads .contextkit/context.md
  • Includes all standards files (code-style, testing, architecture, guidelines)
  • Sends full context to your AI tool (Aider, Claude, Gemini)

contextkit ai <prompt>

Same as above, but with explicit "ai" keyword.

contextkit ai "create a button component"

Set your preferred AI tool with export AI_TOOL=aider

AI Tool Commands

Cursor (Auto-loads)

All context files are automatically loaded. Reference specific commands in chat:

@.contextkit/commands/analyze.md

VS Code Copilot

Use @ mentions to include context files in your prompts:

@.contextkit Create a login button

✅ One mention includes all context files

Claude / Gemini CLI

Reference files directly in your prompts:

read .contextkit/standards/code-style.md

When using ContextKit in AI tools, reference these files in your prompts:

@.contextkit/commands/analyze.md

Analyze and customize standards for your project

@.contextkit/commands/create-component.md

Create a new component following your patterns

@.contextkit/commands/create-feature.md

Create a new feature with all necessary files

@.contextkit/commands/run-tests.md

Run the test suite with coverage

@.contextkit/commands/quality-check.md

Run linting, type checking, and quality checks

@.contextkit/commands/add-documentation.md

Add documentation to your code

Short Alias Available

Use ck instead of contextkit for faster typing. Example: ck install

Was this helpful?

Help us improve the documentation by sharing your feedback.