noname-iii/dsh-code-checker

Name: noname-iii/dsh-code-checker
Author: noname-iii

Plugin插件 Native原生 ⭐ 2 GPL-3.0 Docs & Writing文档与写作

Deepseek Harness 的插件（可能也支持其他 AI 代理 IDE，但我尚未测试），用于在 AI 完成代码编写后检查错误。

Project Overview项目介绍

Edited by Zhiwei Agent (zhiwei)由知微Agent(zhiwei)编辑

This plugin is built natively for DeepSeek Harness (DSH) as a primary target, with an official DSH bundle manifest for installation. It runs a three-step comprehensive check after an AI finishes writing code or building a project, collects any detected issues, and feeds them back to the AI for fixing until the check passes with a “No problems” result. It also includes an optional local GUI dashboard, and can be used as a standalone CLI tool or MCP server for other AI agent platforms including Trae, Qoder, Cursor, and Claude Desktop.

The three-step check process starts with build validation: it detects the project type (from Node.js and Python to C++ and static web), installs dependencies, runs all build commands, and collects any errors. If the build passes, it moves on to feature completeness checking, matching user requirements against the implemented code and even simulating real user interaction to verify features work as expected. Only after the first two steps pass does it run automated tests and simulate full real user operation to catch crashes, freezes, or unresponsiveness. This automatic check-fix-recheck loop continues until the project passes, making it useful for any DSH user building software projects.

The plugin is released under the open source MIT license, and requires Node.js 20 or newer to run, with Node 22 LTS recommended to match DSH’s system requirements. Desktop application simulation is only supported on Windows via UIA, and if Playwright is not installed, web simulation falls back to basic HTTP probes without affecting other check steps. LLM analysis is enabled by default which incurs minor token cost, but you can disable it for zero LLM cost, and after installation you can run the included sanity check script to confirm the plugin is working correctly.

这是一款专为深度求索Harness（DSH）打造的三步式代码全面检查插件，自带DSH原生安装清单。AI完成代码编写和项目构建后，它会自动运行完整检查，收集所有问题并反馈给AI修复，直到输出“没有问题”的结果。它还附带可选的GUI控制面板，同时也支持作为独立CLI和MCP服务供Trae、Qoder、Cursor、Claude Desktop等其他平台使用。

检查流程分为三步，首先检测项目类型，安装依赖并执行构建，收集所有构建错误；接着核对用户需求，验证所有功能是否完整实现，还会通过Playwright或HTTP探针模拟真实用户访问验证功能；最后运行自动化测试，模拟真实用户操作软件，记录崩溃、无响应等异常。整个流程会自动循环直到项目通过检查，适合所有在DSH中开发软件项目的用户使用。

该项目采用MIT许可证开源，运行需要先安装Node.js 20及以上版本，推荐使用Node 22 LTS以匹配DSH要求。桌面端模拟仅支持Windows系统，未安装Playwright时网页模拟会降级为HTTP探针，不影响其他步骤运行。默认开启LLM分析会产生少量token消耗，可关闭该选项实现零LLM成本，安装后可通过内置测试脚本验证插件是否正常工作。

View on GitHub在 GitHub 查看

Send to your DSH发给你的 DSH 安装

请帮我安装这个 DSH 插件。安装前先完成【兼容性检查 + 安全性检查】，检查通过再动手。

插件：dsh-code-checker（noname-iii/dsh-code-checker）
仓库：https://github.com/noname-iii/dsh-code-checker
本站详情页：https://www.yhbd.top/plugins/noname-iii-dsh-code-checker/
本站登记：类型 plugin · 归类 原生 DSH 插件 · 许可证 GPL-3.0 · ⭐ 2 · 最近提交 2026-08-20 · 主语言 TypeScript

按下面顺序执行，每步先把结论告诉我，再进入下一步：

【1 兼容性检查】
① 我这边：DSH 版本、Node 版本、操作系统、当前 profile（web / desktop）。
② 读它的 README、package.json、插件 manifest，列出它要求的 DSH 版本 / Node 版本 / 操作系统 / 外部依赖 / 需要另外先装的运行时。
③ 逐条比对，结论只写「满足 / 不满足 / 未知」三种；不满足的给出可行替代方案。
④ 检查是否和我已装的插件冲突：命令名重复、skill / tool 重名、端口占用、重复注册的 MCP server。

【2 安全性检查】
① 仓库可信度：和上面「本站登记」是否一致；star / fork 数、创建时间、最近提交，是否归档或长期停更。
② 安装脚本：逐行看 package.json 的 preinstall / install / postinstall，以及 install.sh、setup.ps1 之类脚本。出现 curl|bash、下载后直接执行、混淆代码、访问与插件功能无关的域名，立刻停下来告诉我，不要继续装。
③ 依赖：列出新增依赖，标出无人维护、或与知名包拼写近似的可疑包（typosquatting）。
④ 权限与副作用：它会读写哪些目录、访问哪些域名、需要哪些 DSH 权限（filesystem / network / shell / clipboard 等），以及怎么卸载和回滚。
⑤ 如果它要求 sudo / 管理员权限，或权限明显超出功能所需，先停下来问我。

【3 安装】
上面两步没有「不满足」和「高危项」时才执行；用官方推荐方式安装，不要自行提权。

【4 汇报】
用表格输出：检查项 / 结论 / 依据 / 是否需要我决策。拿不准的一律写「未知」并说明要我怎么确认——不要猜，也不要替我决定。

Send this message to DSH in your current session: it verifies compatibility and security first (answering met / not met / unknown item by item) and only installs once everything checks out — it will stop and ask you if it finds a high-risk item. The box scrolls; the copy is the full prompt. CLI install commands may not be accurate across systems, so DSH is the safer route.把上面这条消息直接发给当前会话里的 DSH：它会先核对兼容性与安全性（逐条给「满足 / 不满足 / 未知」），确认没问题再安装，有高危项会停下来问你。框内可滚动，复制到的是完整提示词；安装命令不一定准确，发给 DSH 更稳。

Pre-install check安装前体检Compatibility · Security兼容性 · 安全性 1 warning1 项注意

Only 2 stars - very few users, little community feedback星标只有 2，几乎没人在用，遇到问题缺少社区反馈

DSH walks through these 9 checksDSH 会逐条核对这 9 项

Compatibility兼容性

DSH, Node, OS and profile requirementsDSH 版本 / Node 版本 / 操作系统 / profile 是否满足要求
External dependencies and runtimes (Electron / Python / Docker, ...)外部依赖与运行时（Electron / Python / Docker 等）是否齐备
Conflicts with installed plugins: command names, skill / tool names, ports, duplicate MCP registration与已装插件是否冲突：命令名、skill / tool 重名、端口占用、重复 MCP 注册

Security安全性

Repo matches the facts registered here; archived or abandoned?仓库是否与页面登记一致，是否归档或长期停更
Safety of preinstall / install / postinstall and install.sh / setup.ps1preinstall / install / postinstall 与 install.sh、setup.ps1 是否安全
curl|bash, download-then-execute, obfuscation, unrelated domains → stop immediatelycurl|bash、下载即执行、混淆代码、无关域名 → 立刻停止
Typosquatting or unmaintained packages among the new dependencies新增依赖里有没有 typosquatting 或无人维护的包
Requested permissions vs. what the feature actually needs申请了哪些权限、是否超出功能所需（filesystem / network / shell / clipboard）
Any sudo / admin requirement, plus uninstall and rollback是否要求 sudo / 管理员权限，以及卸载与回滚方式

Anything uncertain must be marked unknown with a note on how to confirm it. This site's signal screen is a static snapshot, not a security audit.拿不准的必须标「未知」并说明要我怎么确认。本站的信号筛查是静态快照，不能替代安全审计。

Or use CLI install (for developers)或使用命令行安装（适合开发者）

CLI Install命令行安装

dsh plugin --profile web add dsh-code-checker

把 noname-iii/dsh-code-checker 加入你的 DSH 配置（web profile）即可启用。

READMEREADME

22.0 KB · 缓存于 2026-08-25 04:53:44 · 22.0 KB

dsh-code-checker · Comprehensive Code Check Plugin

A plugin for DeepSeek Harness. After the AI writes code / builds a project, this plugin runs a three-step comprehensive check and reports every problem straight back to the AI so it can fix them, until it returns "没有问题" (No problems). Optional GUI dashboard included, plus a standalone CLI and an MCP server for Trae, Qoder, Cursor, Claude Desktop and any other platform.

See README.zh.md for the full documentation (中文).

What it does

Build & run check — detect the project type (Node/Python/Rust/Go/C++/Java/.NET/static web/Electron/desktop exe), install deps, run ALL build commands, start a run probe, and collect every error. Any error → report the specific error info to the AI immediately (with file:line locations where available) and list ALL collected errors at once (later steps are skipped).
Feature completeness — extract ALL of the user's requirements from the prompt/context, then verify each one against the implementation (heuristic keyword/structure checks + optional LLM deep analysis + behavioral verification: open the project like a real user — web via Playwright-rendered page text, CLI via --help/--version output — and check whether the required feature is actually visible). Missing features are collected and reported all at once; step 3 runs only when step 1 and step 2 both pass.
Real user simulation — first run the project's automated tests (Node projects: pnpm test; non-Node projects: the equivalent test command such as cargo test/go test ./.../pytest/mvn test/dotnet test), recording any test failure as an error; then operate the software like a real user (keyboard, mouse clicks/drags): web apps (and any GUI project such as a DSH plugin panel) via HTTP probes + Playwright, Windows desktop apps via UIA with real input events, CLIs via driven commands — following the user's described features (or README.md). For web/GUI projects the simulator also audits every page it reaches (detecting pages stuck on a URL or stuck on a "loading" indicator) and clicks every button on each page (recording the state/any anomaly after each click). Any project with a GUI (user interface) MUST run the GUI simulation when steps 1–2 pass — a GUI project never falls back to CLI simulation. Freezes, unresponsiveness, errors and crashes are recorded and reported to the AI. If clean → return "没有问题" and let the AI continue.

Showing the opening section of the README — the full document lives in the repository以上为 README 开头摘要，完整文档在仓库内 · View the full README on GitHub →在 GitHub 查看完整 README →

渲染基于仓库 noname-iii/dsh-code-checker 的 README.md。在 GitHub 查看原文 →