ECC — Everything Claude Code

ECC is a production-ready AI coding plugin that acts as an agent harness operating system. It provides 64 specialized agents, 262 skills, 84 slash commands, and automated hook workflows — built from 10+ months of daily multi-harness use.

Works across: Claude Code, OpenCode, Codex, Cursor, Gemini, GitHub Copilot, Zed, and more.

• Repo: github.com/affaan-m/ECC
• License: MIT
• Latest: v2.0.0

Quick Install

Recommended (Claude Code plugin):

# In Claude Code:
/plugin marketplace add https://github.com/affaan-m/ECC
/plugin install ecc@ecc

# Then install rules (plugins can't auto-distribute rules):
./install.sh --profile core --target claude

Manual full install:

git clone https://github.com/affaan-m/ECC
cd ECC
./install.sh --profile full --target claude

For OpenCode / Codex / Cursor:

./install.sh --target opencode
./install.sh --target codex
./install.sh --target cursor

npm:

npm install ecc-universal
npx ecc install --profile core --target claude

Key Agents

Agent	Role
planner	Complex feature planning and decomposition
architect	System architecture design decisions
tdd-guide	Test-driven development workflow
code-reviewer	Code quality review with inline suggestions
security-reviewer	Security-focused code audit (AgentShield)
build-error-resolver	Diagnose and fix build failures
refactor-cleaner	Code refactoring and cleanup
doc-updater	Documentation generation and updates
loop-operator	Multi-step task orchestration

Plus language-specific reviewers for TypeScript, Python, Go, Rust, Java, C++, Kotlin, Django, and PyTorch.

Essential Slash Commands

/plan              # Complex feature planning
/code-review       # Code quality review
/build-fix         # Fix build errors
/tdd               # Test-driven development
/security-scan     # AgentShield security audit
/model-route       # Route tasks by model complexity
/harness-audit     # Audit harness reliability
/loop-start        # Multi-step task loop

Hooks System

ECC installs hooks for lifecycle events:

Hook Event	Purpose
SessionStart	Initialize context, load project rules
PreToolUse	Validate tool calls before execution
PostToolUse	Log results, detect anomalies
Stop	Session cleanup, save learnings

Continuous Learning (Instinct)

ECC v2 auto-learns from sessions:

/instinct-status             # View learned patterns
/instinct-import <file>      # Import external patterns
/instinct-export             # Export learned behaviors

Patterns evolve from sessions into reusable skills with confidence scoring — no manual curation needed over time.

Token Optimization

For best performance with ECC:

// ~/.claude/settings.json
{
  "model": "sonnet",
  "MAX_THINKING_TOKENS": 10000,
  "CLAUDE_AUTOCOMPACT_PCT_OVERRIDE": 50
}

ECC agents are designed to work within compact contexts — the goal is deep work, not context bloat.