Everything Claude Code

Everything Claude Code is best understood as a workflow pack for Claude Code, not a single feature. It gives you reusable skills and agents for each stage of development: setup, research, planning, testing, review, verification, and cleanup.

Core Workflow

The default mental model is:

Research -> Plan -> TDD -> Review -> Verify -> Commit

That sequence matters more than any individual command. The plugin is opinionated about a few things:

• research before building custom solutions
• tests before implementation
• language-specific best practices
• review and verification before shipping

What It Covers

1. Project setup

• /everything-claude-code:configure-ecc installs skills and rules at the user or project level.

2. Research and planning

• /everything-claude-code:search-first looks for existing libraries, patterns, and prior art before you write code.
• /everything-claude-code:plan turns a vague request into a concrete implementation plan and pauses for confirmation before changes.

3. Implementation and testing

• /everything-claude-code:tdd-workflow enforces a test-first loop.
• Language-specific skills such as /everything-claude-code:frontend-patterns, /everything-claude-code:backend-patterns, /everything-claude-code:python-patterns, and /everything-claude-code:golang-patterns push the implementation toward idiomatic code.

4. Review and verification

• Review agents such as /everything-claude-code:code-reviewer, /everything-claude-code:security-review, /everything-claude-code:python-review, and /everything-claude-code:go-review catch quality and security issues after changes.
• /everything-claude-code:verification-loop runs the final validation pass: build, lint, tests, coverage, and diff review.

5. Maintenance and learning

• /everything-claude-code:refactor-cleaner removes dead code and unused dependencies.
• /everything-claude-code:doc-updater refreshes docs and codemaps after meaningful changes.
• /everything-claude-code:continuous-learning and /everything-claude-code:continuous-learning-v2 capture repeatable patterns from real sessions.
• /everything-claude-code:security-scan checks Claude configuration for risky settings or injection problems.

Scenario 1: Starting a New Project

If I were bootstrapping a project with Everything Claude Code, I would usually follow this order:

1. /everything-claude-code:configure-ecc
2. /everything-claude-code:search-first
3. /everything-claude-code:plan
4. /everything-claude-code:tdd-workflow
5. /everything-claude-code:security-review
6. /everything-claude-code:verification-loop

In practice, that means:

• install the right skills and rules
• research existing solutions before building from scratch
• agree on a plan before touching code
• write tests first
• review for quality and security
• verify everything before commit or deploy

Scenario 2: Refactoring an Existing Project

For a refactor, the workflow shifts slightly:

1. /everything-claude-code:plan
2. /everything-claude-code:refactor-cleaner
3. Apply the relevant language or framework pattern skill
4. Run the appropriate review agent
5. /everything-claude-code:verification-loop
6. /everything-claude-code:doc-updater

A good example stack mapping looks like this:

• React or Next.js: /everything-claude-code:frontend-patterns
• Node.js services: /everything-claude-code:backend-patterns
• Django apps: /everything-claude-code:django-patterns
• Spring Boot services: /everything-claude-code:springboot-patterns
• Go services: /everything-claude-code:golang-patterns

Takeaway

The real value of Everything Claude Code is not that it gives you more commands. It gives you a repeatable engineering loop with stronger defaults: research first, test first, review aggressively, and verify before shipping.

You can invoke the skills explicitly with the /everything-claude-code:* prefix, but many of the patterns are most useful when they become your default way of working.