Easy Coding Agents — Productivity skill for Claude Code
One of the early China-built lightweight Claude Code-inspired terminal coding agents, with autonomous execution, tool calling, custom agents, task control, and evidence-gated memory.
How to install Easy Coding Agents
This entry records only its repository, not the path inside it, so there is no
exact command to give. Open yushui2022/easy-coding-agents and copy the folder into
~/.claude/skills/, or the file into ~/.claude/agents/.
What Easy Coding Agents does
One of the early China-built lightweight Claude Code-inspired terminal coding agents, with autonomous execution, tool calling, custom agents, task control, and evidence-gated memory.
Alternatives in Productivity
- Ccplugins — Claude Code Plugins that actually save time 2.7k ★
- Pi Subagents — Claude Code like Sub-Agents & Workflow Orchestration for Pi — parallel execution, live widget, fleet view, cus 967 ★
- /build-hook - Build Custom Claude Code Hooks — Interactive guide for creating custom hooks with Q&A workflow and automatic validation 618 ★
README
ECA — Easy Coding Agent
**ECA Desktop 0.34.1 Windows 预览版**:桌面工程位于 `desktop/`,支持多项目与多会话并发、主对话与侧聊附件、网页预览、PowerShell 交互终端,以及按需打开的修改审查。与 CLI 共用 Python 核心和项目数据。[安装、开发与验收说明](docs/desktop/README.md)。npm 启动器同时提供 `eca` 和兼容命令 `ecs`。
新 Agent 接手请先读 [完整接力文档](HANDOFF.md):包含本机路径、未提交代码状态、依赖、G 盘缓存、打包更新、测试与陶土界面约定。
主 Agent 默认不限响应轮数和输出截断恢复次数,可用 Ctrl+C 中断。旧配置中的显式上限仍有效,可运行 `ecs --max-turns 0 --output-recovery-limit 0` 取消这两项限制。模型单次输出上限、子 Agent 限额和共享预算的区别见 [输出与持续执行配置](docs/OUTPUT_LIMITS.md)。
你的项目,你的模型。
通过终端或桌面理解代码、执行修改,并检查每一步留下的结果。
CLI + Windows Desktop · v0.34.1 · 预览版
ECA(Easy Coding Agent)是一个可自行配置模型的编程助手,提供 CLI 和 Windows 桌面入口,共用 Python 核心。它可以读取与搜索项目、编辑文件、运行命令,并把完整会话和文件修改记录保存在项目本地。桌面安装包自带运行时;CLI 运行需要 Python 3.11+。
它重视三件事:**改动可检查、工作可继续、记忆有来源。** 你可以先看差异再允许写入,中断后恢复会话,也可以检查或撤销文件工具产生的修改。
当前适合个人试用、开发与研究。基础机制已有自动化测试,但真实复杂任务的稳定性仍在验证中;尚未达到成熟编码 CLI 的功能覆盖与可靠性。
大仓库定位:`/related SYMBOL` 查看定义与 Python 调用/导入候选;模型的 `code_search`、`repo_map`、`code_relations` 支持目录筛选与增量读取统计。[用法与验证边界](docs/REPOSITORY_NAVIGATION.md)。
可选多模型协作
默认只需一个 API,原有用法不变。需要时额外配置云端或本地 Worker,并通过 `worker_profiles` 显式开启,主 Agent 才能按子任务选择模型。不会自动下载本地模型;具体小模型需要支持工具调用并经过任务验证。[配置示例](docs/UPGRADE-v0.13.md)
还可以显式启用 [Worker → 检查 → 主模型修复链路](docs/UPGRADE-v0.14.md):`--worker-first` 指定先执行的模型,只有检查失败且配置修复预算时才调用主模型接手。
能做什么
| 能力 | 实际行为 |
|---|---|
| 终端编码 | 流式输出;读取、搜索、编辑文件;执行需要确认的 Shell 命令。 |
| 本地 Skills(源码新增) | 项目/用户 SKILL.md,/skills、/skill 显式调用;技能不绕过授权。说明。 |
| LSP 重命名(源码新增) | 跨文件预览、冲突检查、审批和修改后诊断;逐文件记录撤销。[边界](docs/SKILLS_REFACTOR_C |
Related Skills
Claude Orgtree
A custom orchestrator for Claude Code that organizes agents into an authority hierarchy. This differs from typ
Verify Loop Skill
A Claude skill that runs a supervised, self-verifying loop on a coding task. The loop is the easy part — verif
Kutu Android Tv Guide
Claude Code + ADB workflow for optimizing older Android TV devices, adding AirPlay screen mirroring and buildi
Harness Improver
Stage 09 at close, or any time mid-engagement. Invoke the harness-improver to review the harness for staleness
Claude Code Dashboard
A lightweight localhost dashboard that monitors multiple Claude Code sessions in real-time. See token usage, c
Mneme Plugin
Persistent cross-project memory and phase-workflow engine for Claude Code — human-gated, local-first, event-so
Related Agents
Vercel AI SDK Expert
Use for Vercel AI SDK v5 implementation work: chat interfaces, tool calling, streaming, agentic loop control,
10x
10x is a terminal-based CLI coding agent (installed via npm install -g 10x-cli) built around the observation t
Hva Desk
The HVA Desk — reads Meta insights by impression volume, runs the deterministic CLEAR scorer, and converts the