// Agent profile
ECC (affaan-m/ECC) is an open-source agent harness optimization system that gives coding agents a coordinated engineering workflow: plan, test, implement, review, verify, remember, and improve. With 245K+ GitHub stars it is the most-starred agent optimization project on GitHub. ECC ships 68 specialized agents, 286 reusable skills, 94 command shims, selective rule packs, runtime hooks for enforcement, continuous learning with instincts, and AgentShield security scanning. It installs as a Claude Code plugin (ecc@ecc), a native Codex marketplace plugin, or via managed adapters for Cursor, OpenCode, Gemini, Zed, Antigravity, Qwen, Hermes, OpenClaw, Kimi, and more. Key workflows include TDD with gated RED-GREEN-REFACTOR evidence, fresh-context code review, build repair, session persistence, and Plan Canvas for visual plan review. ECC Pro adds a GitHub App for private repos. MIT-licensed.
ECC is an open-source agent harness optimization system by affaan-m that transforms how coding agents work. Rather than relying on ad-hoc prompting, ECC installs a coordinated engineering workflow: plan before building, verify changes with tests, review from a fresh context, remember what matters, and turn repeated wins into reusable skills. With 245K+ GitHub stars, it is the most-starred project in the agent optimization category. ECC installs as a Claude Code plugin, a native Codex marketplace plugin, or via managed adapters for 10+ additional harnesses.
ECC ships 68 specialized agents (planning, review, build repair, security, architecture, domain work), 286 reusable skills (TDD, research, security, docs, frontend, data, ML, operations), 94 command shims, selective rule packs by language, and runtime hooks for enforcement. AgentShield provides security scanning across prompts, hooks, MCP configuration, permissions, and secrets. The Unified Memory Vault enables cross-harness context sharing — Claude, Codex, Kimi, and other harnesses read from one local, inspectable Markdown format. Plan Canvas lets you review agent plans visually in a browser with annotation, approval, and Mermaid diagram rendering.
ECC fits any team using AI coding agents that wants a structured engineering workflow instead of freeform prompting. The TDD workflow gates RED-GREEN-REFACTOR with evidence at each step. Fresh-context code review catches regressions and blind spots that the writing context misses. Build repair automates diagnosis and fix cycles. Session persistence and continuous learning extract patterns from real sessions into reusable instincts. Teams running multiple harnesses benefit from the Unified Memory Vault for seamless context handoffs.
ECC’s full catalog (68 agents, 286 skills) can consume context window budget — selective and minimal install profiles are available for constrained environments. ECC Pro (GitHub App for private repos) starts at $19/seat/month; the open-source core is MIT-licensed and free forever. Shell-backed optional features are not portable to every OS. The plugin advertises its catalog to the model, so teams should use selective profiles when context footprint matters.
Builder.io's open-source framework for building agent-native applications — shared actions, SQL-backed state, identity, tools, skills, jobs, observability, and UI surfaces that all work together.
CLI orchestration tool that installs and configures platform connectors for 16+ social and web sources — Twitter/X, Reddit, YouTube, GitHub, Bilibili, and more. Zero API fees.
Persistent memory layer for AI coding agents — benchmark-backed (95.2% on LongMemEval-S), 92% fewer tokens per session vs full-context pasting, zero manual memory.add() calls.