About this project

ECC transforms AI agents from simple code writers into coordinated engineering systems. It implements a structured development lifecycle: plan -> test -> implement -> review -> verify -> remember -> improve. Instead of repeating these processes in prompts, ECC integrates them into the agent's operational framework. Key capabilities include: - Extensive Library: Access to 68 specialized agents (planning, security, architecture, etc.), 286 skills (TDD, research, ML, operations), and 94 command shims. - Runtime Features: Includes hooks for enforcement, session summaries, continuous learning, and context controls. - AgentShield: Integrated security scanning for prompts, hooks, MCP configurations, permissions, and secrets. - Broad Platform Support: Primary support for Claude Code and Codex, with adapters for Cursor, OpenCode, Gemini, Zed, GitHub Copilot, Antigravity, Qwen, and others. - Model Agnostic: Works with official providers, custom API endpoints, or self-hosted models via the harness's native configuration.