Introduction to ECC
Relevant source files
The following files were used as context for generating this wiki page:
Everything Claude Code (ECC) is a production-ready AI coding plugin and ecosystem designed to enhance software development workflows. It provides a comprehensive suite of specialized agents, skills, commands, and automated hooks that integrate with AI coding harnesses to facilitate domain-specific tasks, enforce coding standards, and automate repetitive processes.
The system operates on five core principles: being agent-first by delegating tasks to specialized sub-agents, following a test-driven development (TDD) approach with a minimum 80% coverage requirement, maintaining a security-first posture, enforcing data immutability, and requiring detailed implementation planning before code execution.
Sources: AGENTS.md:1-12
System Architecture and Components
ECC is organized into several distinct layers that handle different aspects of the development lifecycle. This modular structure allows for high cohesion and low coupling across the various automated tools.
Core Components Table
| Component | Quantity | Purpose |
|---|---|---|
| Agents | 67 | Specialized sub-agents invoked via the Task tool for domain-specific operations. |
| Skills | 279 | Canonical workflow surface containing domain knowledge and best practices. |
| Commands | 92 | User-invoked slash commands for specific workflows (e.g., /plan, /code-review). |
| Hooks | Multiple | Trigger-based automations (PreToolUse, PostToolUse, etc.) for safety and validation. |
| Rules | Multiple | Always-follow guidelines for common coding styles and per-language standards. |
| Sources: AGENTS.md:3-5, AGENTS.md:143-151, CONTRIBUTING.md:104-120 |
Repository Structure
The project directory is structured to separate logic types, ensuring that workflows (skills) are distinguished from delegation roles (agents) and automated triggers (hooks).
agents/ — Specialized subagent role prompts
skills/ — Reusable workflow skills and domain playbooks
commands/ — Maintained slash-command compatibility shims
hooks/ — Hook behavior and safety gates
rules/ — Language and harness rules
manifests/ — Selective install modules and profiles
docs/ — Architecture notes and guides
Sources: skills/ecc-guide/references/summary.md:21-31, AGENTS.md:143-151
Workflow and Logic Flow
The ECC ecosystem prioritizes a "Skills-first" approach, where skills serve as the primary workflow surface. However, complex tasks often involve the orchestration of multiple agents.
Implementation Workflow
The standard development cycle in ECC follows a structured path from planning to completion:
The diagram shows the sequence from initial planning to final commit using specialized ECC agents.
Sources: AGENTS.md:122-132, CONTRIBUTING.md:273-286
Agent Orchestration
Agents are invoked based on the specific needs of a task. For example, a request for a complex feature automatically triggers the planner, while a bug fix triggers the tdd-guide.
The sequence diagram illustrates how different ECC agents interact to fulfill a development request.
Sources: AGENTS.md:39-51, AGENTS.md:103-107
Core Engineering Rules
ECC enforces strict coding standards through its rules engine to ensure maintainability and security.
Immutability
A critical principle in ECC is the total avoidance of mutation. Objects and arrays must never be modified in-place; instead, new copies must be returned.
Sources: rules/common/coding-style.md:3-10
Security-First Guidelines
Before any commit is finalized, the system (via agents or hooks) validates the following:
- No Hardcoded Secrets: Prevention of API key or password leaks.
- Input Validation: All user inputs must be validated at system boundaries.
- Vulnerability Prevention: Checks for SQL injection, XSS, and CSRF protection.
Sources: AGENTS.md:54-63, examples/CLAUDE.md:41-47
Testing Standards
- Minimum Coverage: 80% across the codebase.
- Test Types: Unit tests (utilities/components), Integration tests (API/DB), and E2E tests (Critical flows).
- TDD Workflow: RED (Write failing test) -> GREEN (Write minimal pass) -> IMPROVE (Refactor and verify coverage).
Sources: AGENTS.md:103-116
Navigation and Discovery
ECC provides internal tools to help users navigate its large surface area of agents and skills.
Key Discovery Commands
| Command | Usage | Purpose |
|---|---|---|
/ecc-guide | /ecc-guide <topic> | Navigates agents, skills, and docs from the live repo. |
/project-init | /project-init | Stack-aware onboarding for a target project. |
/harness-audit | /harness-audit | Deterministic repo readiness scoring. |
/security-scan | /security-scan | Review configuration security. |
| Sources: commands/ecc-guide.md:1-15, commands/ecc-guide.md:65-71 |
Conclusion
Introduction to ECC reveals a highly structured environment that moves AI-assisted coding from simple completions to sophisticated, agent-driven engineering. By enforcing TDD, immutability, and security through a vast library of modular skills and specialized agents, ECC aims to provide a reliable framework for building enterprise-grade software with AI.
Sources: AGENTS.md:1-12, skills/ecc-guide/references/summary.md:5-17