Everything Claude Code
github.com/affaan-m/everything-claude-code- Category
- AI Agents
- Rank
- No. 25Tools index
- Pricing
- Open Source
- Type
- TOOL
- Builder
- affaan-m
- GitHub
- 254.5k stars
- Latest release
- v2.2.1
- Date
About
A comprehensive performance optimization system for AI coding agents that provides skills, memory management, security scanning, and research-first development patterns. Works across multiple AI agent harnesses including Claude Code, Codex, Cursor, and OpenCode with production-ready configurations evolved from 10+ months of intensive daily use.
What it does
Everything Claude Code installs a reusable engineering workflow around coding agents. It guides work through planning, tests, implementation, independent review, verification, and retained context. Plain Markdown artifacts carry requirements and plans between sessions, while optional hooks, rules, specialist agents, commands, and local memory add structure around that core loop.
Why it's ranked here
The strongest case is breadth joined to an explicit workflow, not merely a large prompt collection. Installers track managed files, planning artifacts remain readable and versionable, and memory handoffs treat recalled material as unverified context. The catch is uneven harness support and enough overlapping surfaces to demand careful setup.
What's good
Its planning model produces human-readable, diffable artifacts that survive session resets and can travel with code. Managed uninstall records protect unrelated user configuration. Memory entries are create-only, scoped, and explicitly treated as unreviewed. Security scanning covers prompts, hooks, permissions, secrets, agent configuration, and tool-server configuration.
Tradeoffs
Claude Code receives the fullest experience, while other adapters intentionally omit capabilities whose runtime contracts are unverified. Users must choose one installation method per harness or risk duplicate skills, commands, hooks, and configuration. Plugin installs cannot distribute rule packs. The repository also combines a broad agent toolkit with a separate provider-agnostic Python package, increasing conceptual surface area.
How to use it well
Use it when a team wants repeatable, auditable agent work across substantial features, especially where requirements, plans, tests, and handoffs should persist beyond one chat. Start with a selective profile and add only relevant rules. It does not replace provider credentials, external service authentication, static security tools, or governed project documentation.
Technical notes+
README.md defines the staged plan, test, implement, review, verify, remember workflow and warns against stacking installation methods. package.json publishes ecc-universal, includes harness-specific directories, installers, hooks, rules, MCP configuration, agents, commands, and a large skills catalog. docs/QWEN-GUIDE.md and docs/JOYCODE-GUIDE.md describe state-file-based managed installs and selective removal. docs/HERMES-SETUP.md documents the ecc memory CLI plus an opt-in ecc-memory-mcp stdio server with per-harness identities and scoped access. pyproject.toml separately packages src/llm as llm-abstraction for Python 3.11+, exposes the llm-select CLI, and declares Anthropic and OpenAI dependencies; src/llm/providers/__init__.py also exports adapters for Ollama, Atlas, and Astraflow.
Observed
- License
- MIT
- Distribution
- Public npm package ecc-universal, Claude Code plugin, GitHub App, shell installer, PowerShell installer, and manual repository setup
- Interfaces
- Agent-harness plugins and adapters, ecc CLI, llm-select CLI, browser dashboard, and optional stdio memory MCP server
- Harness support
- Best support for Claude Code, supported Codex sync, and capability-limited adapters for Cursor, OpenCode, Gemini, Zed, GitHub Copilot, Antigravity, Qwen, JoyCode, and others
- Languages
- Shell, TypeScript, Python, Go, Java, Perl, and Markdown are represented
- Python packaging
- Hatchling wheel package requiring Python 3.11 or newer, with Anthropic and OpenAI runtime dependencies
- Install safety
- Managed adapters record installed-file ownership and remove only ECC-managed files during uninstall
Read from README.md, package.json, pyproject.toml, src/llm/__init__.py, src/llm/__main__.py, src/llm/core/__init__.py, src/llm/tools/__init__.py, src/llm/prompt/__init__.py, src/llm/providers/__init__.py, docs/QWEN-GUIDE.md, docs/HERMES-SETUP.md, docs/JOYCODE-GUIDE.md, docs/TROUBLESHOOTING.md, docs/PLAN-PRD-PATTERN.md.
What it can do
Optimize AI coding agent performance
AI coding agent configuration → Optimized agent performance settings
Manage memory for AI coding agents
Agent memory data and usage patterns → Optimized memory allocation and management
Scan code for security vulnerabilities
Source code files → Security vulnerability report
Provide skills training for AI agents
Agent capabilities and training data → Enhanced agent skills and abilities
Enable continuous learning for AI agents
Agent performance data and feedback → Updated agent learning models
Configure production-ready AI agent settings
Development environment parameters → Production deployment configuration
Integrate with multiple AI coding platforms
Platform-specific agent configurations → Unified cross-platform agent setup
Tags
Tech Stack
Media
Comments (0)
No comments yet
Editorially curated, with community endorsements as a secondary signal. Corrections welcome.