TrendWhat rose on GitHub, every day at 07:30 KST · 한국어

GitHub · as of November 8, 2026 · View on GitHub →

affaan-m/ECC

AI coding agent optimization system

+4,944 stars in 7 days · trending 25 of the last 30 days

Build an engineering system where coding agents plan, verify, and remember

ECC automates the process of agents planning before work, verifying changes with tests, and converting recurring success patterns into reusable skills. It embeds this engineering workflow as the agent's default behavior, eliminating the need to input complex instructions for every prompt and establishing a systematic environment where agents plan and review their own work.

Redesigning the agent's workflow

ECC guides agents to follow a cycle of 'plan -> test -> implement -> review -> verify -> remember -> improve.' This enforces a method where agents verify changes with tests, re-review their work in new contexts, and retain important information. Once installed, this process becomes the agent's default behavior without requiring explicit instructions in every prompt. It operates on the principle of optimizing the context window while maintaining all other state permanently.

Included agents, skills, and security components

This repository provides a large number of specialized agents and skills for various development scenarios.

  • AgentShield security scanning It includes built-in functionality to scan prompts, hooks, MCP configurations, permissions, secrets, and agent files to detect security issues.
  • Hooks and memory system It provides hooks and memory features that handle runtime enforcement, session summaries, persistent learning, instincts, and context control.

Installation paths and supported environments

ECC works best with Claude Code and supports guided installation paths for Codex and Kimi Code. It provides adapters with limited functionality for other agents or editors such as Cursor, OpenCode, Gemini, Zed, and GitHub Copilot. When using multiple agents simultaneously, you must select only one installation method per agent to avoid issues caused by duplicate installations. You must check the support status matrix before assuming feature parity for a specific agent.

Pre-application checks and license

ECC is open source distributed under the MIT license, allowing for personal and commercial use. However, it does not provide identical features across all agents, so you must verify the support status of the specific agent (harness) you intend to use. Some agents, such as Cursor or Gemini, use an adapter approach rather than native plugin support, which may limit hook or native skill discovery features. Before installing, you must install only from official channels (the GitHub repository, npm packages ecc-universal, ecc-agentshield) and avoid third-party re-uploads or unofficial mirrors, which may contain malicious code. Additionally, note that using multi-* commands requires a separate ccg-workflow runtime installation.

Written by AI from this repository's README on October 11, 2026. For the facts, the original on GitHub is the reference.

By the numbers

GitHub trending history
72 days between Jun 19, when Trend began recording, and Nov 8, 2026 · 25 of the last 30 days · first on Jun 23 · best daily rank #1 (Sep 6)

Read from GitHub, as of November 8, 2026.

Language
JavaScript
Topics
ai-agents · anthropic · claude · claude-code · developer-tools · llm · mcp · productivity
Stars
299,556
Stars gained in 7 days
+4,944
Contributors
478
Forks
44,259
Latest release
v2.2.3 · October 1, 2026
Commits
4,327
Last commit
November 6, 2026
Open issues
94
Open pull requests
150

Related repositories

View on GitHub → · Homepage

README badge

trend.ondo.sh · GitHub daily trending #1 · September 6, 2026

The badge above shows the highest this repository has ranked on GitHub's daily trending since June 19, 2026, and the day it did. Paste the Markdown below into your README. The badge opens this page.

[![trend.ondo.sh](https://trend.ondo.sh/github/r/affaan-m--ECC/badge.en.svg)](https://trend.ondo.sh/github/r/affaan-m--ECC/en.html)