The performance optimization system for AI agent harnesses. Skills, instincts, memory, continuous learning, security scanning, and research-first development. Production-ready from day one.
Not configs. A complete system evolved over 10+ months of intensive daily use building real products.
65 skills and 40 commands covering TDD, code review, security scanning, E2E testing, refactoring, and multi-agent orchestration.
Auto-extract patterns from sessions into reusable instincts. Confidence scoring, import/export, evolution across projects.
Hooks that save and load context across sessions automatically. Your agent remembers what worked and what didn't.
Checkpoint and continuous evals, grader types, pass@k metrics. Quality gates that catch regressions before they ship.
Git worktrees, cascade method, model routing by complexity and budget. Scale instances when the task demands it.
1,282 tests, 102 rules. Detect secrets, prompt injection, unsafe patterns. Run security scans directly from your harness.
Plugin install or manual clone. Your choice.
# Option 1: Plugin (Claude Code) /plugin marketplace add affaan-m/everything-claude-code /plugin install everything-claude-code@everything-claude-code # Option 2: Clone + Install git clone https://github.com/web3-claw/claude-code.git cd claude-code ./install.sh typescript # or python, golang, swift, php # Option 3: npm (OpenCode / Codex) npm install ecc-universal
First plugin to maximize every major AI coding tool.
| Feature | Claude Code | Cursor | Codex | OpenCode |
|---|---|---|---|---|
| Agents | 16 | Shared | Shared | 12 |
| Commands | 40 | Shared | Instruction | 31 |
| Skills | 65 | Shared | 10 | 37 |
| Hook Events | 8 | 15 | — | 11 |
| Rules | 34 | 34 | Instruction | 13 |
| MCP Servers | 14 | Shared | 4 | Full |
| Security | Hooks | Hooks | Sandbox | Hooks |
Production-ready agent harness. Open source. Free forever.