开源开发者 CLI 工具，已验证可构建和运行

git 命令的简单终端 UI

测试发现：lazygit builds, all 32 unit-test packages pass, all integration tests pass headlessly, and the TUI binary launches and runs on the virtual display. 12 分钟。

开源 AI 求职 agent 和职位查找工具：扫描职位公告板，在申请前根据你的简历对每个职位进行 1-5 分评分，定制符合 ATS 的简历和求职信，并提供面试准备和申请跟踪器。它帮助你填写每个申请表；你只需按下 Submit。可在本地的 Codex、OpenCode 和其他 AI CLI 中运行，或作为 Claude...

测试发现：All core tests pass on Node 22 with --experimental-sqlite. Playwright Chromium 151.0.7922.34 is installed and launches. The doctor and update-system checks work. The one remaining failure is in web/tests/lib/apply-cv-resolver.test.mjs which needs --experimental-strip-types for .ts imports. 66 分钟。

ripgrep 会递归搜索目录中的正则表达式模式，同时遵循你的 gitignore。

测试发现：ripgrep 15.2.0 built from source, all 332 tests pass, binary searches correctly. 3 分钟。

带翅膀的 cat(1) 克隆

测试发现：bat 0.26.1 builds from source, all 293 tests pass, and the binary runs and syntax-highlights input correctly. 5 分钟。

你的 coding agent 所依赖的运行时

测试发现：herdr 0.8.2 builds, runs its server, passes all Rust unit tests, Python maintenance tests, Bun integration tests, and plugin marketplace tests with no failures. 19 分钟。

GNU coreutils 的跨平台 Rust 重写

测试发现：uutils coreutils 0.11.0 builds cleanly with Rust 1.98.1 on Ubuntu 24.04, produces a working multi-call binary (verified: --version, --help, cat, sort, echo, true, false, ls, head, md5sum, wc), and passes 5078 of 5129 functional tests, all 51 failures are container-environment issues (no TTY colors, no /dev devices... 24 分钟。

正确的终端字符串样式

测试发现：chalk v6.0.0 installs, builds, imports as ESM, and passes all 58 tests with Node 22. The npm registry entry validates as published. 4 分钟。

一个用于构建终端用户界面（TUI）的 Rust crate

测试发现：The Ratatui workspace builds cleanly and all 3176 tests pass on Rust 1.98.1 (stable); interactive examples require a real TTY which is unavailable in the container. 8 分钟。

为什么它在运行？追踪任意进程、端口、容器或文件，追溯到启动它的源头：CLI + TUI。

测试发现：witr builds, passes all 8 test packages (including -race), and traces real processes in the container. 4 分钟。

n³ 非传统终端文件管理器

测试发现：nnn v5.3 builds from source and runs as a terminal file manager, successfully listing directories, navigating into subdirectories, and returning; the test harness confirms all assertions pass. 8 分钟。

你的 CLI 家庭视频录像机

测试发现：VHS builds, all 4 Go packages pass tests (lexer, parser, token, main), and runs end-to-end against a real auto-downloaded Chromium browser + real ttyd + real ffmpeg, producing valid GIF and text output files. 16 分钟。

一个用于代码结构化搜索、 lint 和重写的 CLI 工具，使用 Rust 编写。

测试发现：ast-grep 0.45.3 installed via npm, pip, and cargo build from source; all 202 unit/integration tests pass; search, scan, rewrite, and test subcommands verified working on real code. 9 分钟。

又一个跨平台的图形化进程/系统监视器。

测试发现：bottom 0.14.9 builds from source, all 342 tests pass, and the btm binary responds to --version and --help. 5 分钟。

美观且高度可定制的 Claude Code CLI 状态栏，支持 powerline、主题等。

测试发现：Dependencies installed, project builds to dist/ccstatusline.js, all 1897 tests pass, and piped JSON input renders a formatted status line with model name and git info. 6 分钟。

一种查看和导航目录树的新方式。

测试发现：The Rust project broot v1.59.0 builds from source with cargo build, the broot binary runs and reports version and help, and all 132 tests pass (125 unit tests, 7 integration tests, 0 failures). 3 分钟。

团队中的每个 coding agent 共享一个记忆。Claude Code、Cursor 和 Codex 获得你的代码和团队过往决策的地图，从而从你停下的地方继续。

测试发现：npm install completes, TypeScript compiles, the CLI builds/checks/queries a structural code graph on a real repo without any keys, and all 1220 tests run with 0 failures. 10 分钟。

为团队中的每个 coding agent 提供一份共享记忆。Claude Code、Cursor 和 Codex 会获得你代码和团队过往决策的地图，从而从你停下的地方继续。

测试发现：The graft CLI installs, builds, and its full test suite passes (1491/1492, 1 deliberate skip) with both previously-failing shim-resolution and lock-release tests green, and the CLI builds/parses/queries a scratch repo and completes an LLM deep pass against a local mock endpoint. 87 分钟。

简化创建美观且可测试的命令行界面的过程

测试发现：Symfony Console component library installed with composer dependencies; PHP 8.4.26 runs all 2036 unit tests; 1997 pass, 25 skipped, 17 known environmental issues unrelated to the library code. 33 分钟。

列出你系统中的所有 node_modules 目录及其占用空间。然后你可以选择要删除哪些目录来释放空间。

测试发现：npm install completes, TypeScript compiles cleanly, 236/236 unit tests pass, and the npkill CLI binary runs (version 0.12.2, help text, and JSON scan mode all verified). 2 分钟。

Bubble Tea 的 TUI 组件

测试发现：The project compiles and all tests pass using the standard Go toolchain with Charm's Go 1.26.0 bootstrap. 7 分钟。

markdown 终端幻灯片工具

测试发现：presenterm 0.16.1 builds, all 520 unit tests pass, non-interactive commands (--help, --version, --current-theme, --list-comment-commands, --acknowledgements, --export-html) function correctly; HTML export of demo presentation produces valid 324KB output. 7 分钟。

用于 WebSockets 的命令行客户端，类似于 netcat（或 curl）用于 ws://，具备类似 socat 的高级功能。

测试发现：websocat 1.14.1 compiles from source, binary serves WebSocket connections and passes all 7 integration tests (trivial, tcp, ws, ws_ll, ws_persist, unix, abstract_). 2 分钟。

CLI 工具和 Python 库，可将流行的命令行工具、文件类型和常见字符串的输出转换为 JSON、YAML 或字典。这使得输出可以管道传输到 jq 等工具，并简化自动化脚本。

测试发现：jc v1.25.7 is installed in a Python 3.12 venv at /tmp/jc_venv, all 1557 non-skipped tests pass, and the CLI parses command output to JSON correctly. 1 分钟。

一个带有 Python 绑定的命令行工具和 Rust 库，用于从用户提供的测试用例生成正则表达式。

测试发现：grex 1.4.6 builds from source, its CLI prints version 1.4.6 and generates correct regexes from stdin input, all 693 Rust tests and 23 Python tests pass, and the Python extension wheel installs and builds regexes via RegExpBuilder. 4 分钟。

友好且快速的 HTTP 请求发送工具

测试发现：xh 0.26.2 builds with cargo, cargo test passes 281 tests (1 ignored: expired test certificate), and the installed binary answered HTTP requests against a local Python server (GET 200 with JSON body, POST echo with JSON object, --curl translation). 5 分钟。

当你真的只想马上通过 HTTP 提供一些文件时。

测试发现：miniserve 0.35.0 builds from source, runs all 276 integration tests across 16 modules, serves HTTP on localhost, and generates valid tar/zip archives with correct content. 67 分钟。

用于管理电子邮件的 CLI

测试发现：Rust toolchain 1.99.0 installed, himalaya v2.1.0 built from source with all default backends (imap, smtp, jmap, gmail, msgraph, maildir, mbox, sieve, pimdir), binary executes and reports version, full test suite of 161 tests passes cleanly. 3 分钟。

直观的查找与替换 CLI（sed 替代工具）

测试发现：sd CLI version 1.0.0 builds, passes all tests, and correctly performs find-and-replace operations from stdin and in-place file modification. 2 分钟。

一个免费的 coding agent，运行在你的终端中。它运行在当下最智能的免费模型上。

测试发现：The ANUS CLI installs cleanly, all 28 node --test tests pass, and the --version flag returns 0.2.2. 2 分钟。

一个命令行工具，用于生成、分析、转换和处理颜色。

测试发现：pastel v0.12.0 built from source, all 74 tests pass, binary executes and correctly converts colors between formats. 2 分钟。

一个能了解你的 agent 在做什么的终端窗口管理器。支持平铺窗格、工作区、重启后依然保留的会话，以及每个 coding agent 一个 Inbox。官方渠道仅限 tuios.dev 和本 repo。tuios 没有 X/Twitter 账号。

测试发现：tuios and tuios-web binaries build from source, the full Go test suite (49 packages) passes, and the daemon creates sessions, accepts typed input via send-text/send-keys, and captures pane output showing shell command execution. 22 分钟。

一种新型进度条，具备实时吞吐量、ETA 和非常酷的动画。

测试发现：alive-progress 3.3.0 is installed in a venv, passes all 563 tests, and the progress bar renders correctly via alive_bar. 1 分钟。

一个现代的 watch 命令。时间机器和分页器等。

测试发现：viddy builds in debug and release profiles, all 35 unit tests pass, the help and version commands respond, and the TUI launches and runs under a pseudo-terminal (script) without panicking or errors. 10 分钟。

面向 TypeScript 和 JavaScript 的代码库智能分析。从单一图谱中识别健康状况、复杂度热点、重复代码、架构边界、循环依赖、设计系统漂移以及未使用代码。支持 CLI、GitHub Action、LSP、MCP 和 VS Code。基于 Rust 开发，采用 MIT 许可证。

测试发现：Fallow CLI installs via npm and cargo build, runs on this repo producing valid analysis output, and passes 17018 tests across all workspace crates. 58 分钟。

用于终端的电影化 Git commit 回放工具，将你的 Git 历史变成生动、动画化的故事。

测试发现：The gitlogue binary builds from source, all 228 unit/integration tests pass, and the CLI responds with correct help text and version 0.11.0. 4 分钟。

一个 Ruby gem，用于美化终端的 ls 命令，带有颜色和 font-awesome 图标。:tada:

测试发现：colorls gem 1.5.0 installed from source, all 40 integration tests and 90/91 RSpec tests pass, binary lists directory contents with colorized icons. 19 分钟。

一个用户友好的 SQL 数据库 TUI。使用 Python 编写。支持 SQL Server、MySQL、PostgreSQL、SQLite、Turso 等。

测试发现：sqlit is installed in a venv at /home/runner/.venv, the CLI and TUI launch correctly, the SQLite test suite passes with 2018 passing tests, and CLI queries against a real SQLite file return correct JSON and CSV output. 18 分钟。

一个在终端上绘制图表的命令行工具。

测试发现：YouPlot 0.5.0 installed and working: uplot bar/hist/line/scatter/density/boxplot all produce terminal plots; 109 unit tests pass with 0 failures. 10 分钟。

tmux 的会话管理器，基于 libtmux。

测试发现：tmuxp 1.74.0 is installed in a Python venv at ~/venv, tmux 3.4 extracted from Debian packages runs via LD_LIBRARY_PATH and PATH, all 770 tests pass (2 skipped: suppress_history test and AI-generated skip), tmuxp load creates sessions from YAML workspaces, tmuxp freeze snapshots them, and debug-info reports all... 11 分钟。

一个更好用且友好的 vi(vim) 模式插件，适用于 ZSH。

测试发现：The zsh-vi-mode ZSH plugin v0.12.0 sources cleanly in an interactive ZSH 5.9 with ZLE modules, creating vi-mode keymaps with 95 normal-mode, 43 insert-mode, 19 visual-mode, and 13 operator-pending-mode bindings, plus 103 zvm_* functions for all documented features. 6 分钟。

你的 agent 在终端中，配备本地工具：编写代码、使用终端、浏览网页。在此基础上打造你自己的持久自主 agent。

测试发现：gptme 0.34.0 installed in a Poetry venv; `gptme --version` returns 0.34.0+unknown; `gptme-doctor` reports 22 checks passed, system operational; 1746 out of 1746 core tests pass. 34 分钟。

类似 htop，但用于 AI 编码 agent。实时监控 Claude Code 和 Codex CLI 会话、tokens、上下文窗口、速率限制和端口。

测试发现：abtop 0.5.5 builds and passes all 222 tests; --once and --json modes produce correct live terminal-monitor snapshots from local process/file state. 9 分钟。

无需离开终端即可查看.docx 文件的内容。快速、安全、智能，无需 Office。

测试发现：doxx 0.1.4 builds and runs successfully on Ubuntu 24.04 with Rust 1.99.0. All 144 tests pass. The binary exports .docx files to text, markdown, JSON, CSV, and ANSI formats correctly. NO_COLOR=1 env var must be cleared for color-dependent tests. 9 分钟。

curl 的强大功能，httpie 的易用性。

测试发现：curlie builds from source, passes its args tests, and makes real HTTP requests to the internet returning correct responses. 17 分钟。

终端文件管理器

测试发现：walk v1.13.0 builds from source, passes its 2 Go tests, answers --version and --help, and launches as a TUI under a pseudo-terminal. 22 分钟。

用 Go 构建的极速 TUI 下载管理器，面向高级用户

测试发现：Surge builds from source and all 26 test packages pass cleanly with zero failures. 21 分钟。

使用 Claude Code 将复杂代码库转化为清晰、可导航的架构图。

测试发现：The oh-my-mermaid package installed, built, passed all 37 tests, and all CLI commands (init, list, status, config, read, tree, validate, diff, refs, setup) produce correct output. 1 分钟。

一个快速且强大的日志查看器和处理器，能够将 JSON 日志或 logfmt 日志转换为清晰的人类可读格式。

测试发现：The hl binary v0.36.3 builds from source, all 1796 unit tests pass, and the tool correctly parses and renders JSON log files into human-readable output with working field filtering and level filtering. 6 分钟。

原生支持 iTerm 和 Kitty 的终端图片查看器

测试发现：viu 1.6.1 builds, installs, and runs; 1 unit test passes; binary displays images as ANSI terminal output. 19 分钟。

文件管理自动化工具。

测试发现：The project installs with pip, its full test suite of 266 tests passes, and the CLI runs and reports v3.3.0. 2 分钟。

Go 软件包，用于在命令行应用中生成轻量级 ASCII 折线图 ╭┈╯，且没有其他依赖项。

测试发现：Go 1.22 extracted from debs; asciigraph library compiles and all 27 test groups pass; the CLI binary /tmp/asciigraph builds and renders line graphs from stdin data. 7 分钟。

一个简单的命令行记事本

测试发现：Dnote CLI and server compile from source, their test suites pass entirely, and the server starts serving HTTP on port 3001 with a working /health endpoint. 12 分钟。

获取 npm 包源码，为 AI coding agents 提供更深入的上下文。

测试发现：The opensrc CLI and docs site both install, build, and run. CLI fetches source from npm, PyPI, and crates.io and provides cached paths for grep/search. Docs site builds and serves on port 3456 returning HTTP 200. Rust compilation tests fail due to missing cargo (no root), but the pre-built native binary is downloaded... 13 分钟。

由 AI 驱动的 README 文件生成器。

测试发现：readmeai installs in a venv, passes its full 373-test suite, and generates README output files via the CLI in offline mode. The pyproject.toml dependency bounds were relaxed from caret to minimum versions to avoid Rust-compiler requirements for newer prebuilt wheels. 5 分钟。

使用 fast.com 测试你的下载和上传速度。

测试发现：npm install and tsc build succeed; CLI help and download speed measurement work end-to-end against fast.com; upload speed tests time out due to fast.com rate-limiting from the container IP. 56 分钟。

一个采用模块化架构的 Python 工具，用于可视化并强制依赖项，开源，可通过 pip 安装，可增量采用，实现时无运行时影响，可与现有系统互操作，使用 rust 编写。

测试发现：tach 0.35.1 is fully installed and working: pip install from PyPI succeeds, the Rust extension loads, dependency checking passes on the repo itself, and both the full Python test suite (88/88) and Rust test suite (70/70) pass. 8 分钟。

开发与生产环境的进程管理器

测试发现：dekit v0.10.0 builds from source on x86_64 Linux; all 359 tests pass and the CLI (help, up --help) produces valid output. 26 分钟。

Interactive Redis：一个用于 Redis 的终端客户端，具有自动补全和语法高亮功能。

测试发现：IRedis installed, all 791 tests pass against a real Redis 7.2 server, and non-interactive CLI produces correct output (PING -> PONG). 7 分钟。

自学习 agent 浏览器

测试发现：Webcmd 0.8.4 is installed, builds from source, runs unit tests (3571 passed), connects to Cloak browser daemon, and responds to CLI commands. 5 分钟。

终端界面，用于使用本地媒体播放器查找、下载和流式播放电影、电视剧和电视直播。

测试发现：MovieBox-TUI 0.1.26 builds from source on Linux x86_64, all 394 unit tests pass, and the binary responds correctly to --help and --version flags. 21 分钟。

用于 Go CLI 应用程序的多进度条

测试发现：Built and installed Go 1.26.0, compiled the project, and ran 250 tests, all passing. 5 分钟。

终端中的 VS Code

测试发现：Build succeeds with Node 22, all 118 tests pass, and `tode --serve` starts a code-server instance that responds to HTTP requests. 7 分钟。

本地优先的 AI token 用量与成本追踪器，支持 31 款编码工具，包括 Claude Code、Codex、Cursor、Gemini 和 DeepSeek Harness，并提供原生应用。绝不读取提示词。

测试发现：All 3052 CLI tests pass with 0 failures; the tokentracker serve command starts and returns HTTP 200 on port 7680. 26 分钟。

现代 C++ 的表格生成器

测试发现：tabulate builds with CMake from source, all 18 tests pass (empty_rows, format_propagation, nested_tables, column_operations, alignment_and_wrapping, trim_mode, iterators, border_corner_styles, unicode_borders with single_header variants), and sample binaries produce correct table output. 2 分钟。

在终端中显示丰富的 git commit 图，如同魔法

测试发现：Rust 1.99.0 installed, dependencies fetched, project compiles with cargo build, all 94 unit tests pass, binary prints version and help. TUI cannot initialize without /dev/tty (expected in headless CI). 5 分钟。

终端里的 Hacker News

测试发现：circumflex (clx) builds, runs, passes all 29 test suites, and responds to --version (5.1-dev) and help commands. 6 分钟。

CLI 搜索与替换 | Space Age seD

测试发现：sad v0.4.32 builds from source with cargo, its 2 unit tests pass, and the binary performs batch file edits: piping a file path to 'sad PATTERN REPLACE --commit' writes the edited file in-place. 2 分钟。

功能丰富的终端文本查看器。它所谓的终端分页器。

测试发现：The ov pager compiles, all unit tests pass, the binary runs and responds to --version, --help, --generate-config, and pipe input via --quit-if-one-screen. 6 分钟。

免费 macOS CLI，用于清理磁盘空间：缓存、日志、Homebrew、Xcode。开源替代 CleanMyMac。

测试发现：Installation and build succeed with pnpm + Node 22. All 344 tests pass across 40 files. CLI runs correctly: --version (1.3.5), --help prints full usage, scan finds cleanable files, categories lists all categories. 6 分钟。

用于常见 GraphQL 开发工作流的命令行工具

测试发现：graphql-cli v5.0.0 builds all 6 core packages, the CLI binary works (--help, --version, discover all succeed), and the integration test passes using real generate and codegen commands against a local schema project. 15 分钟。

一个多 shell 补全二进制文件。

测试发现：carapace binary installed at /home/runner/go/bin/carapace, version 'carapace-bin develop', reports 12,587 completers, all 13 test cases pass, formatting/lint/staticcheck all clean, shell snippets and MCP server functional. 12 分钟。

用于 HTTP、GraphQL、gRPC、WebSocket 和 SSE 的 Terminal API 客户端。Plain.http 文件可进行 diff 和版本管理，并支持脚本化测试、工作流、mock、录制器，以及用于 CI 的 CLI 运行器。

测试发现：resterm builds from source with Go 1.26.8, all 64 test packages pass, the CLI runner executes real HTTP requests (verified 204 from httpbin.org), the mock server serves responses (verified 200 on /health), and the init command creates a new workspace. 21 分钟。

在终端中体验 Agentic 聊天。使用自然语言构建应用程序。

测试发现：The Amazon Q CLI builds, passes its 328-unit test suite (0 failures, 13 pre-existing ignores), and the chat_cli binary prints its help text with all subcommands. 6 分钟。

从容管理多个 git 仓库

测试发现：gita v0.16.8.2 is installed in a Python 3.12 venv, all 72 tests pass, and the CLI correctly lists repos and displays multi-repo status. 5 分钟。

构建美观的 PHP CLI 菜单，简单而强大，富有表现力的领域特定语言。

测试发现：PHP 8.5.8 is installed at /home/runner/php, Composer 2.10.3 installed, 67 dependencies resolved via composer install, and the full PHPUnit test suite (454 tests, 949 assertions) passes cleanly. 5 分钟。

用于 AI Agents 的语义搜索与调用图（100% 本地）

测试发现：grepai binary builds and runs; all 16 test packages pass with zero failures. 5 分钟。

一个快速的 codemod 工具部分替代品。通过基于正则表达式的查找和替换，协助大规模代码库重构，并有人工监督和偶尔干预。

测试发现：fastmod v0.4.5 builds, passes all 15 tests, and runs correctly as a find-and-replace tool from the command line and via 'cargo install'. 7 分钟。

通过 RDAP、DNS 和 WHOIS 进行批量域名可用性检查，支持前缀/后缀排列组合、正则表达式模式、MCP 以及多种输出格式

测试发现：tldx builds from source, its 294-test suite passes, and the binary answers version/help, performs real RDAP availability checks, and runs its MCP server over stdio without any third-party credentials. 32 分钟。

极简 AI coding agent（约 1,000 行 Python），受 Claude Code 启发。可与任何 LLM 配合使用。可以将其视为 coding agents 领域的 NanoGPT。前身为 NanoCoder。

测试发现：CoreCoder installed in a virtual environment; 185 tests all green; CLI --help and --demo mode both work; ruff reports one pre-existing EXE001 lint on examples/plan_hooks_demo.py. 1 分钟。

用于被动 OSINT 检查的开源引擎：21 个工具，以 CLI、MCP 服务器和 Web UI 形式提供。托管 Monitor & Diligence 位于 app.openosint.tech。

测试发现：OpenOSINT 2.29.0 installs from source, CLI shows help with 20 tools, DNS tool resolves real domains, playbook generates real investigation reports, web server serves on port 9877 with HTTP 200 on / and /api/health, and 712 automated tests pass. 7 分钟。

Sharing 是一个命令行工具，用于从 CLI 向 iOS 和 Android 设备共享目录和文件，无需额外的 client 应用。

测试发现：npm install completed cleanly, all 32 tests pass, the server starts on a specified port and serves HTTP content with 200 OK. 1 分钟。

从文件设置环境变量

测试发现：env-cmd installs, builds (dist/ pre-built), all 127 tests pass, CLI outputs --version (11.0.0), --help, and correctly loads and injects environment variables from env files into child processes. 7 分钟。

一个用 Rust 编写的快速、独立的终端音乐播放器：原生 Spotify 流媒体，以及本地、Subsonic、电台、Qobuz 和 Youtube 源。

测试发现：spotatui 0.41.0 builds, passes 1282 tests, and runs (--help, --version) in a container with no root access, after installing Rust via rustup and extracting libasound2-dev from its .deb. 18 分钟。

开源策略引擎，用于阻止不良 AWS 支出模式进入生产环境，并修复已产生的超额支出。

测试发现：CloudBurn monorepo installs, builds, lints, typechecks, and passes all 1804 unit tests, 46 e2e tests, and 13 package-install tests across 5 packages; the CLI binary runs and correctly scans Terraform/CloudFormation fixtures. 5 分钟。

通过安装前审计，安全地安装 npm 软件包

测试发现：npq is installed with Node 24, all 777 tests pass, the CLI runs and audits packages via OSV/Snyk returning JSON output with real security findings. 2 分钟。

轻松管理你的 GnuPG 密钥。

测试发现：gpg-tui 0.11.2 builds from source, answers --version (gpg-tui 0.11.2) and --help, and all 21 unit tests pass with a real GPG keyring. 35 分钟。

响应你、你的模型、你的机器和你的规则的 coding agent。

测试发现：Zero v0.9.0 builds from source, passes all 87 test packages, passes go vet, format check, vulncheck, smoke test, and answers all CLI commands (help, models list, providers list, doctor) on linux/amd64. 9 分钟。

响应你、你的模型、你的机器、你的规则的 coding agent。

测试发现：Zero builds from source, all 87 test targets pass, release build produces version 0.9.0 and passes smoke test, and the CLI launches and responds to commands. 9 分钟。

SSH 终端用户界面

测试发现：sshs builds from source, all 34 unit tests pass, and the TUI binary launches on the virtual display printing its version and starting the curses interface. 4 分钟。

用 Rust 编写的快速 Markdown 检查器和格式化工具

测试发现：rumdl v0.2.77 (cargo build --release) and v0.2.78 (npm) both installed. All 13662 tests pass. The binary lints markdown files, detects violations like MD009 trailing spaces, and auto-fixes with --fix. 29 分钟。

基于终端的打字测试

测试发现：ttyper v1.6.0 builds from source, passes all 7 unit tests, installs via cargo install, and responds correctly to --version, --help, and --list-languages. The interactive TUI cannot be exercised in this headless environment, which is expected for a terminal-based app. 3 分钟。

终端颜色渲染库，支持 8/16 色、256 色、RGB 色彩渲染输出，支持 Print/Sprintf 方法，兼容 Windows。GO CLI 控制台颜色渲染工具库，支持 16 色、256 色、RGB 色彩渲染输出，使用类似于 Print/Sprintf，兼容并支持 Windows 环境的色彩渲染。

测试发现：All tests pass and the example demo runs, producing colored terminal output from the command-line color library. 11 分钟。

Ruby/ProgressBar 是一个用于 Ruby 的文本进度条库。

测试发现：ruby-progressbar 1.13.0 builds and passes all 244 RSpec tests on Ruby 3.4.8 (self-compiled from source) in a rootless Ubuntu 24.04 container. 12 分钟。

一个极简的、仅头文件的现代 c++ 库，用于终端小工具

测试发现：The rang header-only library builds with CMake on Linux (GCC 13.3, C++11). All 6 doctest-based test cases pass with 21 assertions. The two standalone tests (colorTest, envTermMissing) also run and exit 0. cmake --install produces correct headers, cmake config, and pkg-config files in the install prefix. 3 分钟。

一个用于构建美观、精雕细琢的 CLI、TUI 和基于文本的 IO 的 Rust crate。

测试发现：iocraft v0.9.1 and iocraft-macros v0.2.4 build and all 195 tests pass; example programs (hello_world, borders, use_output) execute correctly producing the expected terminal output and ANSI-rendered layouts. 27 分钟。

一个快速、极简的目录树查看器，使用 Rust 编写。

测试发现：Rust toolchain installed via rustup, lstr v0.4.0 compiles from source, all 93 tests pass, and the binary runs correctly in classic mode (text, JSON, HTML) with all flags functional. 4 分钟。

一个用 Rust 构建的超快速 CLI 工具，用于解码和编码 JWT

测试发现：jwt-cli builds from source, all 63 tests pass, the binary encodes and decodes JWTs with HS256 correctly. 4 分钟。

通用 AI 上下文生成器。每次对话可节省数千个 token，适用于 Claude Code、Cursor、Copilot、Codex 等。

测试发现：codesight v1.19.0 builds with tsc, all 149 tests pass, CLI scans and generates CODESIGHT.md and wiki output using real detectors (no mocks required). 4 分钟。

专为 AI coding agents 构建的 IDE。

测试发现：npm install and npm run build succeed; all 1957 tests pass (143 files); the app launches in Xvfb under electron . --no-sandbox, prints Provider detection+state init, and stays up until signal-terminated by timeout. 5 分钟。

一个用于终端/控制台的实时绘图工具，数据从 stdin 输入

测试发现：ttyplot 1.7.6 and stresstest binaries build from source, run without crash, and the ttyplot binary renders real-time text-mode plots from piped stdin data via ncurses. 17 分钟。

使用 tcping 对 TCP 端口执行 ping。灵感来自 Linux 的 ping 工具。使用 Go 编写。

测试发现：TCPING builds from source, all 50+ test cases pass, and the binary performs live TCP probes against remote hosts with correct latency/packet-loss reporting. 4 分钟。

用于 Claude Code、Codex、Cursor 和其他 coding agent 的 Agent 技能与专家团队。加载一个技能，或安装领域预设、项目钩子和带测试门的并行工作流。

测试发现：oh-my-agent v13.2.1 builds successfully, the oma CLI binary prints version 13.2.1, and all 4264 tests in 323 test files pass with no failures. 53 分钟。

一个模拟安装操作但实际上不安装任何内容的终端应用

测试发现：Building and testing succeeded; the app runs with a PTY and displays its simulated installation interface with the kernel stage. 5 分钟。

跨平台 cli 应用，用于对字符串执行各种操作

测试发现：The sttr CLI is fully built and functional: go test ./... passes (ok, no test files in 3 packages, processors tests pass cached), the binary answers version, md5, base64, upper, sha256, and json-yaml commands with correct output, and no errors or missing dependencies remain. 4 分钟。

像 IntelliSense，但用于 shell

测试发现：The project builds cleanly, all 207 tests pass (199 unit + 8 doctest), clippy lint passes with no warnings, nightly formatting passes, and the intelli-shell 3.4.5 binary responds to --help and --version. 7 分钟。

一个面向 AI agent 和助手的个人上下文存储，可重用你现有的 coding agent CLI（Codex/Claude/OpenCode），内置 Skills/tools 和桌面 GUI，用于跨 agent 和仓库捕获、搜索和重用项目知识。

测试发现：OpenContext CLI installs, builds UI, runs all 59 Node.js tests (0 fail), starts MCP server via oc mcp, starts web UI via oc ui, and all folder/doc/config CLI commands work on node 18.19.1. Rust tests require cargo (not installed). 4 分钟。

用于 PDF 表单的 Python 库与 CLI。

测试发现：PyPDFForm v5.6.0 installed, all 382 tests pass, CLI and library both functional, no errors. 53 分钟。

极快的 `ls` 替代工具，拥有超能力

测试发现：The lla CLI builds from source, passes all 130 workspace tests, and produces correct output for default, long, JSON, and tree views. 5 分钟。

Kubernetes YAML 到 Terraform HCL 转换器

测试发现：k2tf binary compiles and runs, converting Kubernetes YAML to Terraform HCL; all 4 test suites pass. 3 分钟。

基于 Git 的终端待办/笔记管理器，每个任务对应一个 Markdown 笔记页面。单个二进制文件。

测试发现：dstask builds from source and runs: go test passes all 59 tests, the dstask binary creates/git-syncs/resolves tasks via git backend, and the integration smoke test completes every command without failure. 4 分钟。

Python API / tmux 封装

测试发现：libtmux 0.62.0 is installed in /tmp/venv with tmux 3.4 from extracted Ubuntu package; 1026/1026 tests pass (31 skipped, 1 pre-existing locale-specific failure ignored); Server, Session, Window, Pane APIs work correctly live against tmux. 7 分钟。

一个 Bash CLI 框架，同时也是 Bash 命令运行器。

测试发现：The argc binary (v1.24.0) builds from source, all 233 tests pass, and the framework correctly parses CLI arguments as demonstrated by running an argc-based script with -F --bar=xyz --baz a --baz b v1 v2 producing foo: 1 bar: xyz baz: a b val: v1 v2. 3 分钟。

基于终端的 Java / JVM 进度条

测试发现：The progressbar Java library compiles and its full test suite of 25 tests passes cleanly. 6 分钟。

命令行 XML 和 HTML 美化器及内容提取器

测试发现：xq binary built from source, all 16 project test functions pass, and the binary formats XML/HTML, extracts XPath and CSS queries, outputs JSON, and handles file/stdin input correctly. 5 分钟。

一个相当（简单）的 strace 替代品。

测试发现：The lurk-cli Rust crate builds, cargo test passes, and `target/debug/lurk ls` runs a real ptrace trace of the traced command and prints traced syscalls, exiting 0; the pre-existing inability to attach to external PIDs (EPERM, needs privileged ptrace) remains unchanged. 3 分钟。

面向 AI coding agents 的通用技能安装器和包管理器。一条命令，12+ 个运行时。npx ai-agent-skills

测试发现：All 228 tests pass, validation passes, CLI responds with version 4.3.2 and lists 141 skills across a marketing shelf with 76 mktg collection skills, 7 collections total, and generated docs in sync with the catalog. 14 分钟。

从终端搜索、播放和下载动漫、电影和电视剧，支持英语和巴西葡萄牙语。

测试发现：Go 1.27.1 is installed, the Go module builds cleanly, all 36 test packages pass, and the goanime v1.8.7 binary runs and prints help/version on request. 64 分钟。

即时理解任何代码库。面向代码库的系统智能，为人类和 AI 而构建。

测试发现：The project builds from source, all 2561 tests pass (172 test files), the tree-sitter parser correctly extracts entities and relationships from source code, and the CLI binary responds to help and version queries. 9 分钟。

用于检查和管理监听 localhost 端口的服务的 CLI 工具

测试发现：sonar builds, passes all 154+ tests, and lists local ports on the command line via direct scan. 10 分钟。

codespelunker - 一款理解代码结构并按相关性排序结果的 CLI 代码搜索工具。无需索引，支持 CLI、TUI、MCP 和 HTTP。

测试发现：cs (codespelunker) v3.2.0 builds from source, all 5 test suites pass, console search returns results, and HTTP mode serves 200 on /. 15 分钟。

TTT Editor (Terminal Text Tool)：一个在终端中运行、真正替代 VS Code、Zed 和 Sublime 的工具。一个感觉像 GUI 的 TUI。单二进制，零配置。

测试发现：ttt builds to a 17 MB binary at bin/ttt, the --version and --help flags work, the full Go unit/E2E test suite (30 packages) passes, all 100 functional test files (370 tests, 22 expected failures) pass, and 5 of 7 integration test files pass (2 skipped due to LSP/external-change dependencies). 7 分钟。

Warp 的自定义主题仓库，Warp 是一款用 Rust 构建的极速现代终端。

测试发现：The gen_theme_previews.py script runs without errors against all theme directories and produces correct SVG previews and README.md files containing real theme colors. 2 分钟。

生产就绪的开源终端 coding agent，具有可读、分层的代码：权限规则、操作系统沙箱、MCP、技能、子 agent，以及 Anthropic、OpenAI 兼容、Gemini 或本地模型。

测试发现：eagent 0.1.1 installs, builds, runs (dist/eagent.js --version prints eagent 0.1.1), and passes all 68 production verification tests covering CLI, tools, MCP, sandbox, skills, agents, hooks, plugins, UI rendering, and resilience. 6 分钟。

在 ANSI 终端中使用真彩色绘制图像

测试发现：PIXterm 1.3.3 builds, and the pixterm binary successfully renders PNG/WebP/GIF/TIFF/BMP images and a fetched HTTP image to true-color ANSI output, plus reports version 1.3.3, all tested in this container. 3 分钟。

CodeAgent 是一个开源的 coding agent，常驻在你的终端中。它理解你的仓库，规划多步骤变更，编写和编辑文件，运行命令和测试，并持续迭代直到任务完成，整个过程通过透明、可扩展的工具循环实现。

测试发现：CodeAgent v0.1.0 is installed with all 41 Python dependencies in .venv, the 67-item verify regression suite passes, and the app launches showing its banner and version. 6 分钟。

命令行工具，用于根据你的分支模型显示清晰的 git 图。

测试发现：git-graph v0.8.0 builds from source and correctly visualizes git history graphs in normal, ASCII, and SVG output modes with multiple branching models. 4 分钟。

由 Rust 驱动的极简 Claude Code 状态栏 HUD。

测试发现：best-claude-hud v0.1.11 installs from npm, builds from source with Rust 1.99.0, its 50-unit Rust test suite passes, its 2 npm-node tests pass, and the release binary correctly renders a Claude Code statusline containing model, reasoning effort, workspace directory, git status, and context window usage. 3 分钟。

面向 AI coding agents 的代码库 wiki。记录代码无法表达的内容：决策、流程、不变量、注意事项。

测试发现：The codealmanac package is installed and functional: CLI responds to --help and --version, all 564 pytest tests pass, ruff lint passes, and doctor reports a healthy install with 71 indexed wiki pages. 8 分钟。

超级简单的 PDF 发票

测试发现：InvoicePrinter installed with Ruby 3.3.12 via rbenv, all 37 tests pass, CLI generates valid PDF invoices, and the server API returns correct PDF data on POST /render. 10 分钟。

在终端、VSCode 扩展或通过 Discord 免费使用 Claude Code

测试发现：The FastAPI proxy server installs from source, all 884 tests pass, and the server responds HTTP 200 on both / and /docs endpoints. 5 分钟。

在终端窗格中查看 agent 的 diff，并将评论发送回 Claude Code、Codex、OpenCode 或 Pi。这是一个 herdr 插件。

测试发现：All 1009 tests pass across 8 test suites, fmt-check and clippy pass, release binary builds at 17MB, and the TUI launches via PTY rendering its full interface. 10 分钟。

代码搜索/替换工具

测试发现：amber v0.6.1 builds, all 9 unit tests pass, and both ambs (search) and ambr (replace) commands function correctly on real files including regex mode with named and positional capture groups. 6 分钟。

zsh 的预测性内联建议

测试发现：Built, installed, and ran deja end-to-end: clean build, full race-enabled test suite green (111 passed), import/ping/record/query/fuzzy/empty, daemon socket protocol, and SQLite fallback all working. 5 分钟。

快速、独立的 EPUB 到 Kobo EPUB 转换工具。

测试发现：kepubify, covergen, and seriesmeta binary builds succeed, the full Go test suite passes, and kepubify converts a test EPUB into a KEPUB with Kobo span/HTML markers. 3 分钟。

在 Web 浏览器中运行 kubectl 命令。

测试发现：Gotty server with webkubectl API extensions builds from source and runs, serving the index page at HTTP 200, generating tokens via /api/kube-token and /api/kube-config POST endpoints, and serving static assets (JS bundle, CSS, favicon) all at HTTP 200. 6 分钟。

Git 族谱，清晰梳理。一个用于浏览提交图的 TUI，色彩鲜明，一目了然。

测试发现：The keifu Git commit graph TUI builds, passes all 91 tests, lints and formats cleanly, prints version 0.6.0, and its debug server returns a live screen dump showing the repo's 240-node commit graph. 6 分钟。

拦截 LLM API 流量并在实时终端仪表盘中可视化 token 使用情况。跟踪成本、调试提示并监控 AI 开发会话中的上下文窗口使用情况。

测试发现：Tokentap is installed in a fresh venv, all 38 tests pass, and its proxy starts and correctly routes Anthropic and MiniMax requests to the real upstream APIs (verified by their 401 auth responses) while rejecting unknown paths with 400. 2 分钟。

如果你生活在终端里，你的日志应用

测试发现：The Rust project builds and all 176 unit and integration tests pass across JSON, SQLite, and VJOURNAL backends; the tjournal binary prints its version and help text. 3 分钟。

开源 TypeScript 终端 coding agent，专为 DeepSeek-V4 设计，基于 DeepSeek 强劲的性价比和超低缓存定价，通过工程实现字节稳定前缀和缓存复用 fork，使跨会话记忆与连续自纠错层几乎不增加 token 成本；支持 1M 上下文、Skills/MCP/Hooks，兼容...

测试发现：Build succeeds via tsc, all 1832 tests pass across 191 test files, and the compiled binary runs and prints version dao-code v0.5.9. The app requires a DeepSeek API key for real usage, which is expected and documented behavior. 11 分钟。

一个用于获取 CLI 工具帮助的 CLI 工具

测试发现：The Rust project compiles and all 22 automated tests pass. The halp CLI tool runs and successfully probes commands for --help and --version flags (verified with cat and echo). 10 分钟。

为常见 DevOps 工具提供按需密钥访问的 CLI

测试发现：Summon builds from source and runs: the binary answered --help/--version, all 103 unit tests passed, and it retrieved a secret through the built-in Conjur provider against a local HTTPS mock (authenticate then secret fetch, value printed to the child process), plus the external-provider path returned the fake provider... 5 分钟。

查看你的 coding agent 做了什么以及花费了多少。将每个任务分解为工作步骤：使用的工具、更改的文件、运行的测试、花费的时间和 token。本地优先的仪表板，适用于 Claude Code、Codex、OpenCode 等。无需登录，无需遥测。

测试发现：agentacct 0.12.11 installed in ~/.agentacct venv, all 4888 tests pass, demo command produces run reports, local JSON API returns 200 on /. 50 分钟。

捕获并修复 AI coding agents 遗留的代码质量问题，包括死代码、不安全类型转换、吞没的错误、重复代码、安全风险等。涵盖 10 种语言目标，50+ 条确定性规则，支持 CLI、CI 和 GitHub Actions。运行时无需 LLM。MIT 许可证。

测试发现：Build succeeds, CLI outputs version 0.16.1, full test suite passes (2122/2122, 201 files). 13 分钟。

你的 AI 的项目大脑。让 LLMs 即时获得架构上下文，而不消耗 token。

测试发现：codemap and codemap-mcp build and run with bundled ast-grep, the HTTP API answers 200 on /api/health, the MCP server completes the initialize handshake, and the full Go test suite passes. 11 分钟。

从网站以 PDF/EPUB/CBR/CBZ 格式下载漫画的工具

测试发现：The CLI builds and runs, real manga downloads from mangatown.com produce valid PDF and CBZ files containing 20 decodable JPEG pages each, and the site package's deterministic fixture tests pass, while cmd/gui remains unbuildable for lack of X11/GL headers. 74 分钟。

cmd2：快速构建功能丰富且用户友好的交互式命令行应用程序，使用 Python

测试发现：cmd2 is fully installed and its 1864-test suite passes (2 skipped) with no failures; the documentation builds cleanly, and the quick-start example produces correct output. 4 分钟。

用于时间追踪的命令行工具，采用人类可读的纯文本文件格式。

测试发现：klog builds from source and all 17 Go test packages pass; the CLI parses real .klg files and outputs correct total and report data. 2 分钟。

Spec Driven Development：为 AI coding agents 编写代码库感知的规范，防止它们偏离轨道。

测试发现：Shotgun CLI v0.1.0.dev3 builds, installs, and the full unit test suite (2321 tests) passes with 0 failures after fixing 8 tests affected by the container OPENROUTER_API_KEY environment variable. 55 分钟。

一种用于 JSON、YAML、TOML 及其他序列化格式的路径查询语言。

测试发现：jsongrep 0.10.0 builds from source with cargo, all 263 tests pass, and the jg binary correctly matches JSON path queries on stdin and file inputs. 3 分钟。

面向 AI coding agent 的上下文压缩：在工具输出进入模型之前进行压缩，将重复内容去重为 13-token 引用。支持 Claude Code、Cursor、Codex、Kiro、Zed 以及任意 MCP 客户端。Rust 实现，零 LLM 调用。

测试发现：The sqz workspace builds and its 1715 tests pass; sqz 2.1.1 runs, compresses with byte-exact expand, the sqz-mcp stdio server completes the MCP initialize handshake and lists 7 tools, and the local proxy answered HTTP 200 on POST /v1/chat/completions with a compressed body. 13 分钟。

面向真实产品的自主开发。受管控的 agent 根据你批准的规范进行构建，遵循你设定的规则，并且在另一个 agent 验证通过之前不会合并任何内容。你的决策、判断和品味是工作必须达到的标准。宣告你的愿景，让 agent 去构建它。

测试发现：Pullboard 0.9.0 is installed, the CLI answers commands, the view server binds a port, the spec check reports 0 errors, and the test suite passes 1410 of 1414 tests (1 container-specific parent-death runner test fails, 3 skip without Chrome). 59 分钟。

面向 AI 编码会话的 ~97% token 减少：零依赖、38 种语言、MCP 服务器

测试发现：SigMap v8.51.4 generates deterministic code signature maps successfully, all 246 tests pass (26 unit + 8 R-language + 14 secret-scan + 10 strategy + 174 integration + 14 Python AST), CLI context generation produces 98% token reduction on the repo itself, validate reports 99% coverage, and verify correctly detects... 4 分钟。

减少 Claude Code 在会话启动时加载的 token 数量。一个 CLI 工具，用于设置一个小型的 CLAUDE.md，包含按需文档，测量启动 token，并保持 token 数量较小。

测试发现：All 414 tests pass, the CLI builds and runs all 10 commands correctly, 5 hook scripts fixed to use bash arithmetic instead of bc, Node upgraded from 18 to 20, and the package installs with only engine warnings. 8 分钟。

一个强大的 Git worktree CLI 工具，具有自动设置、分支跟踪和智能导航功能

测试发现：wtp builds, installs, lints, and passes all unit and e2e tests, and its add/list/cd/exec/remove commands were verified against a real local git repository. 5 分钟。

面向 AI coding agents 的终端多路复用器。在同一个 tmux TUI 中并排运行 Claude Code、Codex、OpenCode、Gemini CLI 等，支持实时状态、worktrees 和 diff 审查。

测试发现：agent-manager binary builds from source, all 28 test suites pass, formatting and vet checks are clean, and the binary prints its version and help text successfully. 15 分钟。

适用于 Unix 的 Python 启动器

测试发现：Rust toolchain installed, project builds and installs, all 103 tests pass, and the installed py binary lists Python 3.12, prints version, creates venvs, and auto-uses them via PYTHONPATH-free directory discovery. 3 分钟。

一个极其安全的加密工具，将你的秘密编码为可爱的猫狗声音，使用真实的椭圆曲线密码学并带有趣味伪装。

测试发现：Installed purr from source and verified genkey, list-keys, set-dialect, import-key, and encrypt/decrypt roundtrips (cat and dog dialects) all exit 0 with byte-identical decrypted output, and the full test suite passes 22/22. 7 分钟。

让 AI agent 在终端之间互相发送消息、观察和创建。Claude Code、Codex、Antigravity CLI、Cursor CLI、OpenCode、Kilo、Pi、Kimi

测试发现：hcom v0.7.26 builds and passes all unit tests (2252 passed, 1 flaky ignored) and all CLI smoke tests (28 passed); the binary outputs correct help, version, and status. 4 分钟。

本地主机，随处可用。一条命令即可为你的开发服务器提供公共 HTTPS URL，支持 WebSockets 和热重载。免费，无需账户，运行在 Cloudflare Workers 上。

测试发现：pnpm monorepo installs cleanly, CLI builds to dist/hostc.mjs, all 51 unit/integration tests pass, wrangler dev serves the tunnel API at localhost:8787 creating real tunnels, and the CLI binary reports version 2.0.4 with correct help output. 7 分钟。

用于管理 $GOBIN 中 Go 安装的二进制文件的快速工具：更新、导出/导入以及跨机器迁移工具集

测试发现：Go 1.27.2 installed from upstream tarball, gup binary builds and runs all subcommands, all 14 unit-test packages pass cleanly. 2 分钟。

简单且安全的机密纸质备份

测试发现：paper-age v1.5.0 builds from source, all 42 tests pass, binary produces valid PDFs with scannable QR codes containing Age-encrypted secrets, and the encrypt->QR->scan->decrypt round-trip succeeds. 8 分钟。

你的 AI coding agent 的可靠性取决于其周围的 harness。用 harness-score 在几秒内测量该 harness。

测试发现：Builds with tsup, all 397 vitest tests plus 7 release tests pass, self-scan returns L4, and the CLI binary produces correct JSON and terminal output against real fixtures. 1 分钟。

Claude Code 的本地分析仪表板。无需云端，无需遥测。

测试发现：cc-lens 0.4.1 installs, builds, and runs on Node 22 with all 54 tests passing and the demo dashboard serving HTTP 200 on all routes. 7 分钟。

Go 代码覆盖率转 SVG 矩形树图

测试发现：The go-cover-treemap binary builds and generates valid SVG treemaps from Go cover profiles via both -coverprofile flag and stdin, supporting all documented flags. 4 分钟。

一个持久化的、基于文件的记忆层，为编码 agent（如 Claude Code、Codex 等）赋予项目大脑（持久化的决策、需求、约束），通过零依赖的 CLI 实现。

测试发现：All 50 project tests pass. Every CLI subcommand runs correctly: init scaffolds BRAIN.md and brain directory, CRUD page operations work end-to-end, hooks install/uninstall cleanly, and wiring creates agent config files. Zero external dependencies or credentials required. 4 分钟。

lowfat：精简命令输出，去除噪音，节省 tokens。

测试发现：The lowfat 0.8.1 Rust CLI builds, its full 225-test suite passes, and the binary runs real commands like 'lowfat ls -la' producing filtered output with measured token savings. 5 分钟。

Claude Code hooks + 可安装的插件市场：安全、成本、可观测性、生产力。

测试发现：All 1919 tests pass and the bench tool runs without errors; ruff and uv are installed locally for the format-code plugin tests that require them. 4 分钟。

使用单个命令，在你的 PC 上或云端，在并行的沙箱虚拟机中运行多个 agent。

测试发现：The AgentBox monorepo builds and all 46 package test suites pass (0 failures, 0 cached failures), the agentbox CLI binary responds with correct version/help output, and rsync is available at /home/runner/.npm-global/bin/rsync for the sync-layer tests. 19 分钟。

面向极客的命令行 Todo List / Task Manager

测试发现：Geek-life TUI app builds cleanly and launches under a PTY, rendering the full terminal UI with Projects, Tasks, Dynamic Lists, title bar, and status bar. 4 分钟。

交互式命令行 JMX 客户端

测试发现：jmxterm builds under JDK 21, all 78 unit tests pass, and the uberjar prints usage help when invoked with -h. 2 分钟。

批量重命名文件。

测试发现：All 74 tests pass and the renamer CLI correctly renames files by find/replace patterns on the command line and via API. 4 分钟。

在 CBOR、JSON、MessagePack、TOML 以及 YAML 1.1 和 1.2 之间进行转换

测试发现：Remarshal 2.1.4 is installed in a Python 3.12 venv; all 196 tests pass; CLI converts between JSON, YAML, TOML, CBOR, MessagePack, and Python without errors. 1 分钟。

以 CLI 为先的 AI 编码 agent 编排：并行运行选定的 agent 和模型，比较原始答案，并协调审查或实现工作流。

测试发现：MCO 0.11.0 Python runtime installed in venv, Node wrapper installed, CLI responds to --version and doctor --json, all tests pass. 7 分钟。

它不会修复你的缺陷。它会修复你。这个调试器把问题抛回给你，而不是给出答案。

测试发现：The project installs in a venv, 167 unit tests pass, the CLI shows help/version, and the demo command runs a full scripted debugging session with scoring. 7 分钟。

Claude Code & Codex 本地 token 追踪：状态栏（Codex 业界首创的伪状态栏）、GitHub 风格热力图、多模型成本分析。

测试发现：Installed per install.sh into a private venv; all 416 repo tests pass, and tt --version, tt setup, tt theme, tt status, tt daily, tt weekly, tt monthly, tt sessions, and tt sidebar --once all exit 0 and render live Codex data from this container. 1 分钟。

从命令行控制浏览器的标签页

测试发现：brotab 1.5.0 installed in .venv, all 33 unit tests pass, CLI commands (bt clients, bt windows, bt list) execute without error, bt_mediator starts and shuts down cleanly. 3 分钟。

面向 AI agent 的结构化代码智能：在单个 Rust 二进制文件中提供语义搜索、知识图谱以及内置的 MCP 服务器。让 Claude、Cursor 以及任何 MCP 客户端深入理解你的代码库。

测试发现：Octocode 0.26.3 builds from source with all features, full test suite passes (1589+136), fmt/clippy are clean, and the binary successfully indexed, searched, and served MCP tools over HTTP on a fresh demo repository. 47 分钟。

一个按半字节编辑字节的终端 UI。

测试发现：The heh hex editor builds, passes 5 unit tests, and launches as a TUI app displaying file contents in hex/ASCII with address labels and 12 data label panels. 3 分钟。

一个在 Java 中直接访问 MySQL InnoDB 数据文件的库和命令行工具

测试发现：Build succeeds, all 434 tests pass, and the CLI jar correctly reads and queries MySQL 5.7 InnoDB .ibd files without external dependencies. 6 分钟。

将任务作为可保留、可复用、可优化的步骤流来运行，而非一次性对话。

测试发现：flowtrace CLI builds, installs, passes all 49 tests, serves the web UI on port 3005, and manages traces (validate, lint, show, run lifecycle, step lifecycle, replies, deliverables) via its CLI and REST API. 8 分钟。

可将 LLM token 使用量减少 60-90% 的 CLI 代理。为 Claude Code、Cursor、Copilot、Gemini 提供声明式 YAML 过滤器。用 Go 实现的 rtk 替代方案。

测试发现：snip builds as a static binary, all unit and race tests pass, all 55 inline filter YAML tests pass, and the binary's help/version/run/gain/config commands all produce correct output. 2 分钟。

使用 Claude 构建大型项目。Chief 将你的工作分解为任务，并循环运行 Claude Code 直到完成。

测试发现：Chief binary builds and runs, all 10 test packages pass, CLI responds to --version, --help, and list commands. 15 分钟。

面向现代多核机器的 Gzip 解压缩与随机访问

测试发现：rapidgzip 0.16.0 installed in /tmp/venv, its CLI decompresses gzip input to byte-identical output, and its full test suite passes. 8 分钟。

用于创建常规提交的交互式 CLI。

测试发现：After installing Rust 1.99.1 to ~/.cargo-rust, koji builds with cargo build --locked, all 57 tests pass, the interactive TUI produced a conventional commit in a live PTY, and the README install path cargo install --locked koji installed a working koji 3.4.0 binary. 8 分钟。

开源密钥与环境管理。安全、简单、协作。CLI 和 GitHub 同步。

测试发现：All 5 workspace packages build successfully, 158 tests pass (117 unit + 41 e2e), the Nuxt app serves HTTP 200 on port 3000, and the CLI binary responds to --help. 12 分钟。

在命令行中使用 CircleCI

测试发现：circleci-cli builds from source with Go 1.27.1, the binary reports version and help, and all 7 clikit packages, all ~30 internal packages, and the full acceptance test suite (261s) pass. 27 分钟。

穿透本地开发端口的迷雾

测试发现：Kiri Rust CLI builds, passes all 157 tests, clippy and fmt checks, runs with ANSI-rendered table output, detects processes via /proc, and gracefully reports missing ss/lsof for port listening detection. 5 分钟。

使用你喜欢的文本编辑器重命名文件

测试发现：pipe-rename 1.6.7 builds and passes all 14 tests in a fresh container. 2 分钟。

agent 技能的任务控制

测试发现：LazySkills binary builds from source, passes its full test suite (including race detector), passes all CI checks (gofmt, vet, deadcode), and its non-TUI commands (doctor, find, update, version) execute correctly against both the filesystem and the external skills.sh registry API. 3 分钟。

用于 DSH Plugin Hub 的开源 CLI、模式、解析器和 DSH agent 工具

测试发现：DSH Hub CLI workspace installs, builds, and passes the full test suite (326/326) along with typecheck and package-content/governance checks when run under Node 22.23.4 and pnpm 10.33.0, and the dsh-hub binary answers --version with 0.5.0. 2 分钟。

开源 AI coding agent，专为大型项目和真实世界任务而设计。

测试发现：Plandex server starts on port 8099 with PostgreSQL 16.4 and LiteLLM proxy on port 4000; server tests pass; CLI builds and authenticates against local server; full CRUD API for projects, plans, branches, settings is operational. 18 分钟。

用 rust 编写的极快 git 终端 UI

测试发现：GitUI 0.28.1 builds from source, its full test suite passes (314 tests), and the binary launches as a terminal UI displaying real git repositories on the configured virtual display. 8 分钟。

你唯一需要的速查表

测试发现：cheat.sh HTTP server is running on port 8002 with no-cache mode, serving cheat sheets from 6 fetched upstream repositories across all configured adapters (tldr, cheat, cheat.sheets, learnxiny, rosetta, late.nz). 16 分钟。

更智能的 cd 命令，支持所有主流 shell。

测试发现：zoxide 0.10.0 builds from source, runs without errors, passes all 16 unit tests, and successfully adds and queries directories in its database. 3 分钟。

一个可靠的 coding agent，用于处理复杂的软件工程任务。

测试发现：Reasonix builds as a single Go binary (bin/reasonix), answers --help with full usage including CLI/TUI, web, serve, ACP, and bot subcommands, passes all 147+ test packages with 0 failures, and produces valid JSON diagnostics via doctor --json. 19 分钟。

一个命令行 benchmark 工具

测试发现：hyperfine 1.20.0 builds, all 102 tests pass, and the binary produces correct benchmark timing for a real command (sleep 0.1). 2 分钟。

一个命令行工具，可让你在 App Store 上搜索 iOS、iPadOS、tvOS、watchOS、visionOS 和 macOS 应用，并下载.ipa 或 macOS.pkg 应用软件包。

测试发现：ipatool binary builds successfully, all 15 unit test packages pass, and the search command reaches the live Apple App Store API and returns real results when provided with a keychain passphrase and a minimal mock account entry. 6 分钟。

运行在终端中的现代 API 客户端。

测试发现：Posting 2.10.0 builds and installs with uv sync, its CLI responds to --help, the TUI launches on Xvfb with correct rendered layout, 126 unit tests pass (non-snapshot), and all 194 tests including snapshot comparisons pass when run serially with NO_COLOR unset. 31 分钟。

在浏览器中审查 AI coding agents 的计划、差异和运行中的应用。本地优先，可与任何 agent 配合使用。

测试发现：Built crit binary compiles, launches an HTTP server on 127.0.0.1, serves /api/health returning 200, auto-detects git changes in feature branches, serves embedded frontend, and the full test suite passes all unit and JS tests except one pre-existing intermittent flake in session lazy-threshold ordering. 42 分钟。

用于 AI coding agents 的终端会话管理器。一个 TUI，支持 Claude、Gemini、OpenCode、Codex 等。

测试发现：Agent-deck vdev builds and runs in the container. The binary responds to version, help, and ls commands. 34 of 34 directly testable internal packages pass. The WebSocket terminal bridge functions correctly. Only pre-existing environment-specific test failures: keepalive client-count assertions (container PTY behavior)... 62 分钟。

IDE 风格命令行自动补全

测试发现：Dependencies install, tsc build and lint pass, 202 unit tests pass, 'is doctor' reports all checks green after 'is init bash', 'is complete' returns real spec suggestions, and the interactive inshellisense session on bash launches, renders suggestions, runs commands and exits cleanly; only the fish/zsh e2e variants... 48 分钟。

用于 coding agent 的开源元 harness。定义你自己的工作流、审查者和修复循环。zeroshot 强制执行你配置的流程。

测试发现：Zeroshot installs via npm and builds from source; the CLI version, template list, UI server (200 on /ui/), and run validation all work, with 1607/1608 Rust tests passing in this container. 25 分钟。

漂亮的 vim 风格 powerline，适用于 Claude Code

测试发现：With Node 22.23.3 from ./.node on PATH, npm ci, npm run build, typecheck, lint, and the test suite all succeed (414/414 tests pass), and bin/claude-powerline renders colored powerline and TUI statuslines from piped hook data with exit code 0. 7 分钟。

Terragrunt 是一个灵活的编排工具，允许用 OpenTofu/Terraform 编写的基础设施即代码进行扩展。

测试发现：Terragrunt v1.1.4 builds from source, the `terragrunt --version` and `--help` commands work, all 49 internal/ and pkg/ unit test packages pass, and catalog/scaffold integration tests pass. 13 分钟。

跨供应商消息传递，用于 CLI AI coding agents：让 Claude Code、Codex、Gemini 和 Copilot 在一个团队中相互通信。Bash + SQLite，无守护进程，无框架。

测试发现：agmsg installed to ~/.agents/skills/agmsg/ (version f5a72de), sends/inboxes/history work correctly between 2 agents on the same team through SQLite store, and the test suite runs with all core functional tests passing. 20 分钟。

对你的代码创意进行现有技术检索，这个开发工具已经发布了吗？

测试发现：patent v0.14.0 installed and working: --version returns 'patent 0.14.0', --list-sources enumerates 19 sources, real searches against live package registries return JSON verdicts (e.g. 110 matches for 'markdown to html converter'), and the full test suite passes 375 tests with 0 failures. 16 分钟。

面向命令行的交互式速查表工具

测试发现：navi builds from source, passes all 26 Rust unit tests, and its core cheatsheet query, info, widget generation, tldr integration, and cheatsh integration produce correct output when run with a controlling terminal. 26 分钟。

Squad：适用于任何项目的 AI agent 团队

测试发现：Squad CLI v0.13.1-build.3 builds from source, runs `version` and `--help`, and passes 3454+ tests across 128 test files on Node 22.13.0. 27 分钟。

AI 上下文的 ripgrep：一个零依赖的 C++23 CLI + MCP 服务器，用于编码 agent。无需阅读仓库即可找到所需内容，然后检查你构建的是否符合预期：blast radius、要运行的测试、质量差异。签名比函数体少 74.7% 的字节；每个猜测都被标记，每个损失都公布。带着地图出发。

测试发现：Build succeeds (plain dev build/ and Release build-install/), binary parses the repo source tree and emits deterministic minified XML with ranked symbols via Personalized PageRank, installed to /home/runner/.local/bin/ripwire with skills and hooks, --version/--help work, Python indexing confirms correct function... 21 分钟。

一个轻量级且现代的任务运行器，用于自动化、CI/CD 与 DevOps。

测试发现：Castor v1.8.1 installed as static binary at /home/runner/.local/bin/castor; PHP 8.4.23 with sodium, zip, curl, openssl at /home/runner/php/8.4.23/bin/php and /tmp/php (static-php-cli bulk build); Composer dependencies installed at /work/repo/vendor; zip/ssh/scp binaries available on PATH; test suite runs with 321... 55 分钟。

DevSpace：最快的 Kubernetes 开发者工具，使用 DevSpace 自动化你的部署工作流，并直接在 Kubernetes 内部开发软件。

测试发现：DevSpace CLI builds from source and all unit tests pass; the binary prints help output correctly and is ready for use with a Kubernetes cluster. 16 分钟。

面向 AI coding agents 的预执行防护，在工具调用运行之前阻止破坏性的 Git 和文件系统命令及常见的敏感文件访问尝试，支持 Amp Code，Antigravity CLI，Claude Code，Codex，Cursor，DeepSeek Harness，Devin CLI，GitHub...

测试发现：bun install and bun run build succeed. bun run check passes lint, formatting, typecheck, knip, and duplication checks, with 3673 tests passing out of 3677 total across 213 files. Coverage is 98.91%. All E2E tests pass (packed-runtime, hermes-openclaw, and protection contracts). 15 分钟。

终端中的 Spotify 播放器，功能完全对等

测试发现：spotify_player 0.25.1 builds and passes all 45 tests, clippy (all features + no-features), and cargo fmt; the binary starts on Xvfd and produces the expected OAuth authentication URL on first launch. All requested features (streaming, rodio-backend, media-control, image, notify, fzf, daemon) compile and run without... 15 分钟。

Multi-harness 控制平面，用于 Claude Code、Codex、Cursor 和 OpenCode：跨多个 Claude/Codex 订阅的配额感知轮换、共享线程上下文和跨模型审查。

测试发现：pnpm install completed, pnpm build succeeded (31 packages in 18s), the CLI runs and reports version 3.9.8, the full test suite passes (350 test files, 4692 tests), canary stories pass (6 files, 50 tests), and typecheck passes (61 tasks). 20 分钟。

CLI 代理，可将常见开发命令的 LLM token 消耗降低 60-90%。单个 Rust 二进制文件，零依赖。

测试发现：RTK (Rust Token Killer) v0.42.4 builds from source, passes all 2936 tests with 0 failures, and runs as a working CLI proxy producing token-compressed filtered output for ls, git, cargo, and other commands. 8 分钟。

面向 AI agent 开发者的审查优先终端 diff 查看器

测试发现：Hunk 0.23.0 builds from source with Bun 1.4.2, passes typecheck, runs its CLI (help, version, log, diff commands), and passes 2372/2383 default tests. 1 pre-existing daemon timing test fails, website tests require optional playwright dependencies, and the changeset versioning test requires Node 22+. 37 分钟。

版本化 Codex 指令部署，包含预览、所有权清单、钩子隔离、场景评估和恢复。

测试发现：The codex-instruct.py CLI v0.6.0 installs, deploys overlay/unrestricted/contract prompts into Codex config directories with timestamped backups and manifest transactions, reports accurate multi-field status, and uninstalls with full rollback, verified by 1198 passing tests and a complete end-to-end... 11 分钟。

一个简单、快速且用户友好的 'find' 替代工具

测试发现：fd 10.5.0 builds from source, all 268 tests pass, and the binary correctly searches files by pattern, extension, and glob. 6 分钟。

面向 Python 的精简应用框架。通过简单的 Python API 构建复杂的用户界面。在终端和 Web 浏览器中运行你的应用。

测试发现：Textual 8.2.8 with syntax extras is installed; the full test suite (3459 tests) passes; the demo app and a minimal app both run and render correctly. 14 分钟。

Typer，构建出色的 CLI，易于编码，基于 Python 类型提示。

测试发现：Typer v0.27.2 is installed, all 1372 tests pass, and the CLI tool runs Python scripts as CLI apps. 20 分钟。

网络诊断工具

测试发现：The Trippy Rust workspace (trippy/privacy/traceroute/ping tool) compiles, passes all 807 unit tests, and the CLI binary reports its version. The project is ready for development. 33 分钟。

用 Rust 编写的极速终端文件管理器，基于异步 I/O。

测试发现：Yazi 26.9.1 builds from source with Rust 1.98.1, both the TUI file manager (yazi) and CLI (ya) binaries produce correct version output, the TUI launches and renders the file listing interactively, and the CLI executes subcommands without error. 22 分钟。

GoAccess 是一个实时 web log 分析器和交互式查看器，可在 *nix 系统的终端中运行，或通过浏览器运行。

测试发现：GoAccess v1.11 builds from source and runs. It parses Apache/Nginx Combined log format files into valid JSON output with correct request counts (4 requests, 3 unique visitors, 2xx status codes, browser/user-agent breakdown) and generates a complete self-contained HTML report. The binary at /work/repo/goaccess executes... 14 分钟。

免费、本地工具，用于跟踪 37 个工具和 agent 的 AI 编码 token 使用量与成本（包括 Claude Code、Cursor、Codex、Gemini 等），按模型、项目和任务统计。npx codeburn

测试发现：CodeBurn v0.9.23 builds, passes 3546 of 3567 tests, TypeScript compiles clean, and the CLI prints help and version on demand. 8 分钟。

源代码复制粘贴检测器。支持 220+ 种语言，Rust 引擎，SARIF/HTML/badge 报告器，GitHub Action，面向 AI agent 的 MCP 服务器。

测试发现：jscpd 5.1.1 builds from source via cargo, installs via npm as a prebuilt binary, and runs duplication detection on 224 language formats with 15 reporters; the full Rust test suite passes. 9 分钟。

:stars: 管理你的 shell 命令。

测试发现：Pueue v4.0.4 builds from source, all 186 tests pass, and the daemon-client pair runs correctly: adding tasks, querying status (plain & JSON), viewing logs, killing, cleaning, and waiting all work over the Unix socket. 10 分钟。

一个非常快速、可移植且可扩展的模糊查找器。

测试发现：The television fuzzy finder builds from source with cargo and all 386 tests pass including 125 CLI integration tests exercising the TUI via phantom-test on Xvfb. 20 分钟。

一个可扩展、极简、快速的 TUI 文件管理器

测试发现：xplr 1.1.1 TUI file explorer built from source via cargo build --locked --release; binary responds to --version and --help; all 22 tests pass. 4 分钟。

首个用于 AI 编程的开源 harness 构建器。让 AI 编程变得确定且可重复。

测试发现：Bun 1.4.2 installed, 1282 dependencies installed; all 9 workspace packages' tests pass with 0 failures; server starts on port 3090 and serves Archon web UI (HTTP 200); CLI prints full help output; one import-boundary test expectation updated in packages/cli/src/cli.test.ts for bun 1.4.2 compatibility. 36 分钟。

快速且极简的 3D 查看器。

测试发现：F3D Python bindings (pip) create a rendering engine and produce valid 300x300 PNG images; the F3D CLI binary reads OBJ files and renders them to valid PNG output on the virtual display. 36 分钟。

用于构建交互式提示的 Rust 库

测试发现：inquire 0.9.4 Rust workspace compiles with rustc 1.82.0, all 250 tests pass when NO_COLOR is unset, all 3 workspace crates (inquire, inquire-derive, examples) build successfully. 33 分钟。

Go (golang) 软件包，包含 90 个可配置的终端 spinner/进度指示器。

测试发现：The Go+ compiler v1.24.13 runs on this system, the spinner package builds successfully, and all 17 tests pass including spinner start/stop/reverse/color/update/backspace/line-computation tests. 8 分钟。

音乐搜索下载器。

测试发现：The music-dl Laravel CLI builds and runs under PHP 8.5.8 in this container: version/help output works, the Pest suite passes (36 passed, 3 skipped), and a real song was searched and downloaded to a valid MP3 via the netease source; the composer 'pest' coverage wrapper and tencent/kugou download sources were not... 27 分钟。

通用二进制安装程序

测试发现：UBI Rust workspace builds cleanly, the library crate passes all 169 tests (archive extraction, forge mock tests, platform detection, asset picking), the CLI parses arguments correctly (21/21 unit tests pass), and the CLI downloads and installs binaries via direct --url. Only the GitHub-API-dependent integration tests... 5 分钟。

在代码冲突之前捕获意图冲突。一个构建在 Git 之上的开源 coding agents 协调协议。

测试发现：All 98 tests pass, both binaries (foremerge, fmg) report version 0.5.0, the query benchmark runs successfully, and the doctor command reports a healthy install with an uninitialized database (expected pre-init state). 3 分钟。

从任何设备访问你的终端和 AI agent：手机、平板、桌面。

测试发现：clsh server starts, serves health check (HTTP 200) and static web UI; PTY creation works via node-pty on Linux. 2 分钟。

使用 Rust 编写的高性能 coding agent harness，专为大规模 multi-agent 开发工作流而构建。

测试发现：jcode v0.81.7-dev is built from source, installed via symlinks in ~/.jcode/builds/current and ~/.local/bin, and works correctly with OpenRouter provider. jcode run and jcode repl both complete successfully with real AI responses. 20 分钟。

一个 CLI 工具，可将你的代码库转换为单个 LLM 提示，包含源代码树、提示模板和 token 计数。

测试发现：code2prompt v4.3.0 builds, passes all 113 tests, and the CLI runs correctly, generating formatted prompts from any codebase path. 10 分钟。

在您的机器上或 AI agent 的环境中运行任何进程，就像它是您的 Kubernetes 集群中的一个 pod 一样：真实的环境变量、DNS、网络、流量。

测试发现：mirrord CLI binary and layer build, run, and pass all unit tests and clippy checks on x86_64 Linux with nightly-2026-08-13 Rust toolchain. 30 分钟。

Log 文件导航器

测试发现：Lnav v0.14.0 is running correctly with full functionality; all tested log formats, SQL queries, filters, and headless operations work; source from the repository can be configured with zig cc as the C++ compiler after provisioning development headers manually. 39 分钟。

使用 agents 对任何内容进行逆向工程，从应用行为到原生二进制文件。

测试发现：npm ci installs 281 packages, tsc builds cleanly to dist/, and the full vitest suite runs 353 test files (1756 tests) with zero failures on Node 22.19.0. 7 分钟。

一个由 GPT-5 等 AI 大语言模型驱动的命令行生产力工具，可帮助你更快、更高效地完成任务。

测试发现：ShellGPT 1.5.1 installed in a Python venv; the test suite (29 tests across test_code, test_default, test_roles, test_shell) passes fully with mocked completions; the CLI entry point answers --help and --version. 4 分钟。

用于 App Store Connect API 的快速、可脚本化的 CLI。自动化 TestFlight、构建、提交、签名、分析、截图、订阅等操作。

测试发现：The asc binary builds from source, prints version and help, and all 110+ Go test packages pass with zero failures under ASC_BYPASS_KEYCHAIN=1 (mock/internal HTTP, no real API credentials). 11 分钟。

Coreutils for Windows：安装程序与打包

测试发现：The coreutils multi-call binary builds and runs 79 UNIX utilities on Linux after adapting the Windows-native project to compile without Windows SDK. 27 分钟。

一个公开的动画宠物画廊，适用于 Codex、Claude Code、DeepSeek Harness、Hermes、OpenCode、Gemini CLI 等。

测试发现：All dependencies installed, Next.js app builds and serves pages (200 on /, /en/, /robots.txt, /sitemap.xml), CLI builds/typechecks/installs pets, 537 of 539 tests pass. 24 分钟。

用于管理 Docker 容器和服务的沉浸式终端界面

测试发现：Dockly installs, lints pass, --version and --help work, and the full TUI initializes against a mock Docker daemon with all dockerode API calls responding correctly. 7 分钟。

通过询问代码的功能来查找代码。一个面向 coding agents 的 CLI，使用 Jev 来发现相关文件和源代码上下文。

测试发现：Repository builds, types check, lints clean. All 4 test suites pass (70 pass, 0 fail). End-to-end CLI search works against a mock provider: auth saves credentials, doctor verifies access, search returns file locations and source excerpts. 6 分钟。

面向 React Native 和 Expo 开发者的 AI 驱动图标生成 CLI，使用 OpenAI 最新的模型在几秒内生成精美的应用图标。

测试发现：SnapAI CLI installs, builds (tsc + webpack), and runs; the icon and feature-graphic commands successfully generate valid PNG images when pointed at a mock OpenAI-compatible API endpoint. Gemini model requires a real Google API key. 5 分钟。

Hetzner Cloud 的 CLI

测试发现：hcloud CLI binary builds and runs, printing version 1.69.0-dev and full help output. The complete internal test suite (34 packages) passes with zero failures. 5 分钟。

一个用于在 shell 中与 Atlassian Jira 交互的文本用户界面。

测试发现：The JiraTUI v1.16.0 project installs cleanly with uv sync, the CLI tool (jiratui) responds correctly to version, help, config, and themes commands, and the full test suite passes over 300 tests covering actions (standard and legacy key bindings), models, search, attachments, git, JQL editor, confirmation screens... 34 分钟。

Posix 脚本，用于从你的终端（fzf）或应用启动器（rofi）浏览 youtube 及其他 yt-dlp 支持的网站，并支持可选预览。（支持 bash、zsh 和 dash）

测试发现：yt-x v0.8.7 installed with yt-dlp, fzf, jq, and mock-mpv; performs YouTube searches via yt-dlp, displays search results, auto-selects first result, and correctly invokes the configured media player with the video URL. 7 分钟。

快速、现代的 traceroute，带实时 TUI、每跳统计、ASN/地理位置查询、ECMP 检测和 MPLS 标签解析。比 mtr 更好。

测试发现：The ttl v0.23.0 Rust project builds from source and passes all 345 automated tests (unit, integration, doc-tests), the binary prints version 0.23.0 and help text, but raw network probing cannot run due to container-level CAP_NET_RAW restriction. 12 分钟。

一个 GitHub CLI 扩展，用于生成你的 GitHub 贡献历史的 3D 模型

测试发现：The gh-skyline binary builds, its full test suite passes (go test exit 0, 80 top-level PASS, 0 failures), and against a local schema-correct mock GitHub GraphQL API it generates ASCII art and writes a structurally valid binary STL (458100 triangles, exact size 22905084 bytes) for single-year, range, and --full modes. 19 分钟。

使用原生 Rust 编写的开源 AI agent 框架，从单个二进制文件提供 GUI、CLI、无头模式和 webapp。支持多提供商、MCP、技能、插件和 agent 团队。

测试发现：Rust 1.99.0, Node 22.14.0, and pnpm 10.34.5 installed. thclaws-core 0.138.0 compiles successfully (non-GUI). Both thclaws-cli and thclaws binaries respond with correct version. Frontend dist built. 2669/2676 unit tests pass; 7 pre-existing thclaws_gateway tests fail with PoisonError from reentrant env locking. 16 分钟。

用一个命令更新多个仓库

测试发现：multi-gitter 0.63.1 builds from source with Go 1.27.2; ./... test suite passes (0 failures); the binary answers version/help, and run/status complete successfully against a mock Gerrit over real HTTP, pushing a commit with MultiGitter-Branch/Change-Id footers and reporting the resulting change as Success. 40 分钟。

Autoprompt 是一种编码 agent 技能，可将 agentic 编码任务的失败率降低 45%。

测试发现：autoprompt-skill v2.0.0 CLI installs globally, all 256 core/provider/benchmark tests pass cleanly on Node.js 20.20.2 with Python 3.12 + PyYAML 6.0.3 + jsonschema 4.26.0. 31 分钟。

用于将 AI 搜索、推荐和对话检索集成到 agent 系统和业务系统的开放 CLI。

测试发现：SearchCLI v0.2.1 installed and running on Node.js v20.11.1 with all 55 acceptance tests passing against embedded mock servers. 2 分钟。

一个轻量级终端编码助手，具有类似 Claude Code 的工作流、工具循环和 TUI 架构，专为学习和实验而构建。多语言支持：现提供 TypeScript、Python 和 Rust 实现。

测试发现：MiniCode installs, compiles, passes all 270 tests, and launches in mock mode successfully. 5 分钟。

DeepSeek coding agent 的 harness：持久化 Shell、Ultra 子代理、自动批准、Chrome MCP 和会话遥测

测试发现：Install and setup complete, 570/586 unit tests pass, dscode exec mode works end-to-end with mocked LLM, TUI app needs real TTY: this is a headless-container limitation, the binary itself is healthy. 13 分钟。

一个轻量级 cli，用于运行单一用途的 AI agent。在 TOML 中定义专注的 agent，然后从任何地方触发它们：管道、Git 钩子、cron 或终端。

测试发现：Installed Go 1.25.4, built the axe binary, fixed the ENOTDIR agent auto-discovery bug, and verified the binary end to end locally: 'axe version' prints 1.10.0, the fresh-user init/list/dry-run flow exits 0, agent runs against the local OpenAI-compatible mock return echo output with exit 0, and the full test suite... 17 分钟。

经典 Unix cp(1) 命令的一个显著更快的替代方案，可在极短时间内复制大型文件和目录。

测试发现：fcp 0.2.2 builds and runs in this container: cargo test passes 18/21 (character_device blocked only by missing system expect), --version prints 0.2.2, recursive directory copy and step-echo PTY character-device copy both verified byte-accurate. 8 分钟。

利用此 TUI MCP Client for Ollama 发挥本地 LLM 的强大功能。具备所有核心 MCP 原语（工具、提示词、资源）、agent 模式、多服务器、模型切换、流式响应、人在回路、思考模式、模型参数配置、系统提示词以及保存的偏好设置。

测试发现：The ollmcp CLI installs from source, builds and renders help, and its full test suite passes 305/305 after a one-file Rich word-wrap fix; MCP server add/list/remove works, but live LLM chat cannot be exercised because no Ollama server is available. 6 分钟。

Claude Code CLI 的中文汉化插件，支持 macOS、Linux、WSL、Windows 和新版 native，安装仅需一行命令，更新后自动修复。

测试发现：claude-code-zh-cn plugin v2.17.2 installed with Chinese language settings, 187 spinner verbs, 41 tips, session/notification hooks, and output style; no Claude Code CLI present in container so CLI patch layer skipped. 18 分钟。

关闭终端窗口，而不是 AI 编码会话。一个数字就能让你回去。所有 Claude Code 和 Codex 会话都在一个菜单中：命名、编号、颜色编码。打开、关闭和跳转到任何需要你的会话。

测试发现：Session Kit installed and CLI commands verified; shpool 0.11.0 built from source; all helper binaries deployed to ~/.local/bin/; 31/45 doctor checks pass (remaining non-ok due to container without systemd). 55 分钟。

Laravel 应用的命令行实时监控。

测试发现：The laravel-top package installs with Composer under PHP 8.4.23 and its own test suite passes (19 tests, 33 assertions); phpstan and pint report 13 pre-existing style/analysis issues each and the PHPStan turbo accelerator warning shows the static build cannot load dynamic extensions. 20 分钟。

Agent 优先的 Solana 交易 CLI：18 个 DEX 适配器，12 个 SWQoS TX 落地处理器。支持买入、卖出、lp、狙击、创建池、启动代币，并从你的终端流式接收实时链上事件。

测试发现：npm install --legacy-peer-deps completed (1000 packages), npm run build compiled with zero TypeScript errors, 47 CI-safe tests (registry + unit) pass on Node v22.2.0, CLI binary at dist/cli.js responds to --version and --help with all 22 commands operational. 9 分钟。

预览优先的 Grok Build 指令部署，具有兼容性和 hook 隔离、持久恢复以及分层卸载。

测试发现：The grok-keysmith CLI runs without dependencies under Python 3.12, passes all 140 tests, and its full deploy→status→run→uninstall lifecycle works correctly against an isolated temporary HOME with a mock grok binary for the run subcommand. 5 分钟。

官方的 Linode 命令行界面。

测试发现：The linode-cli installs, builds from its OpenAPI spec, and passes all 233 unit tests; the CLI binary prints help, version, and command listings on demand without requiring real API credentials. 2 分钟。

官方 Model Studio CLI（阿里云百炼 CLI），专为 AI Agent 框架构建，将模型、搜索、多模态和工作流能力暴露为结构化工具调用。

测试发现：workspace installed, builds, all unit/command/dsh/kscli suites pass, bl and kscli launch and report versions, and a real bl text chat round-trip against a schema-correct local DashScope mock returns the mock's assistant message. 17 分钟。

功能丰富的交互式 Jira 命令行。

测试发现：The jira-cli binary is built, runs help/version commands without error, and all unit tests pass. Real Jira API credentials are required for operational commands. 5 分钟。

一个丰富的 GitHub 终端 UI，不会打断你的工作流程。

测试发现：The gh-dash binary builds from source, runs on Linux amd64, prints help and version output, and passes all 243 unit tests across 19 Go packages. 16 分钟。

每个命令行爱好者都值得拥有的数据库客户端。

测试发现：dblab builds from source and all unit tests pass; the interactive TUI app cannot launch because /dev/tty is unavailable in the headless container. 9 分钟。

轻松从终端配置 macOS 安全设置。

测试发现：stronghold v1.5 installs in a Python 3.12 venv, runs -v and -lockdown modes to completion; lockdown mode exercises all configuration steps (firewall, system protection, metadata, daemons, user safety) via mocked macOS commands with exit code 0. 3 分钟。

基于规范、智能体驱动的工作流框架，专为 AI 编码 agent 设计。将请求转化为可验证的目标循环：计划、执行、验证，并在你的仓库中保留持久的规范和证据。兼容 Claude Code、Codex、Gemini、OpenCode 以及普通 CLI。

测试发现：npm install completed without errors (Node 18 engine warnings only); ospec CLI v2.1.0 responds to --version, --help, init, status, change, session commands; release smoke test and content scan both pass; index rebuild tool works; global install via --prefix succeeds. 5 分钟。

用于复杂代码库的唯一 agent harness。项目记忆、规划、执行和已验证的完成，均在 Codex 内实现。

测试发现：LazyCodex plugin (omo 5.0.0-beta.90) is installed, enabled, and verified via the Codex marketplace path with 9/9 repo tests passing, 288/319 plugin tests passing, 27 skills, 21 hooks, and 4 components (rules, comment-checker, lsp, telemetry) responding to CLI commands. 37 分钟。

用于 Azure CLI 的 Azure DevOps 扩展

测试发现：Azure DevOps CLI extension 1.0.8 is installed and its `az devops` help launches, and 354 unit tests plus 8 integration scenarios run green against an in-process mock of the Azure DevOps REST API. 68 分钟。

一个功能丰富的终端 UI 文件传输与资源管理器，支持 SCP/SFTP/FTP/S3/SMB/WebDAV。

测试发现：termscp builds from source (without SMB), its binary responds to --version and --help, and all 335 unit tests pass. 9 分钟。

终端内的真实浏览器

测试发现：pnpm install and pnpm -r build succeed. Node 22 with --experimental-sqlite passes all browser and shared tests (42/43 pixel tests). CLI binary runs and responds to help/version/ls commands. 9 分钟。

smenu 最初是一个轻量级且灵活的终端菜单生成器，但很快发展成为一个功能强大且多功能的 CLI 选择工具，适用于交互式或脚本使用。

测试发现：smenu 1.5.0 binary builds and runs, reads words from stdin/file, displays in scrollable terminal window with ncurses-based cursor/selection, supports keyboard navigation, and outputs selected words on stdout with exit 0. The automated test harness (tests/test.sh) cannot execute because ptylie requires root for... 20 分钟。

21 世纪的 R 控制台

测试发现：radian 0.7.1 installed via pip, uses R 4.3.3 extracted from Ubuntu debs. 15 of 22 automated tests pass. R console launches interactively (test_aaa confirms prompt). CRAN-based features (reticulate Python mode, askpass, renv) untested due to no CRAN mirror access. 30 分钟。

Agent OS：agent 会自行变得更聪明。我们只需守住底线：面试门槛、分阶段评估、预算化进化循环。MCP server，14 个运行时：Claude Code、Codex CLI、Gemini CLI、OpenCode、Copilot、Kiro 等。

测试发现：All unit tests pass (0 failures in ~21000 tests across 34 unit directories), ruff lint and format clean, CLI boots with correct version, MCP server help displays correctly, 2 Windows-specific bugs were caught and fixed. 73 分钟。

Go 项目的工具

测试发现：Go 1.23.4 installed; project builds without error; README generated from page.yaml via go generate; binary runs and exits successfully. 3 分钟。

交互式、文件级、类似 Time Machine 的工具，用于 ZFS/btrfs/nilfs2（甚至 Time Machine 和 Restic 备份）

测试发现：httm 0.51.0 builds and runs on Rust 1.99.0; the binary prints its version, help text, and debug config correctly, but requires ZFS/BTRFS/NILFS2 snapshots for data output. 4 分钟。

Android Linux CLI 启动器

测试发现：Both fdroid-debug and playstore-debug APKs build successfully from source with Gradle 8.2, Android SDK 34, and a self-signed dev keystore. 9 分钟。

CodeMachine 是一个开源工具，用于将 AI 编码 agent 编排成可重复的、长时间运行的工作流。

测试发现：CodeMachine CLI v0.8.0 installs via 'bun install' (591 packages), responds to all CLI subcommands (help, version, auth, run, step, templates, agents, mcp, import, export), and launches its TUI interface. No test suite exists in the repository. 5 分钟。

在你的终端中进行 AI 结对编程：一个静态二进制文件，亚毫秒级启动，支持任意模型。

测试发现：Prebuilt agentty 0.9.26 linux x86_64 release binary is installed at /home/runner/.local/bin/agentty and responds to all query subcommands (version, status, config doctor, skills, help, diagnostics). Source build fails at the jaal submodule due to insufficient compiler support for C++26 structured binding packs. 30 分钟。

用于 Solana 交易的 Agent-first CLI：18 个 DEX 适配器、12 个 SWQoS TX 落地处理器。买入、卖出、提供流动性、狙击、创建池、启动代币，并从你的终端流式传输实时链上事件。

测试发现：无法验证；log 显示了它停在哪里。 21 分钟。

出色的代码覆盖率报告，无需麻烦

测试发现：无法验证；log 显示了它停在哪里。 1 分钟。

Windows Terminal 的一个分支，原生集成 agent，直接在你的命令行中。

测试发现：Rust toolchain installed and working. Installer bootstrap crate builds and its 2 unit tests pass. WTA crate type-checks for Windows MSVC target. C++ component (Visual Studio/MSBuild) cannot build on Linux. Full test suite (1997 WTA tests) requires Windows to execute. 32 分钟。

使用 Rust 编写的开源终端 agent。可使用任何提供商或模型。始终欢迎 Issues 和 PR。

测试发现：无法验证；log 显示了它停在哪里。 19 分钟。

Kyanos 是一个使用 eBPF 的网络分析工具。它可以可视化数据包在内核中花费的时间，捕获请求/响应，使故障排除更高效。

测试发现：无法验证；log 显示了它停在哪里。 38 分钟。

从 TUI 或 Web 管理多个 Claude Code、OpenCode agent，方便在移动设备上访问。还支持 Mistral Vibe、Codex CLI、Gemini CLI、Pi.dev、Copilot CLI、Factory Droid Coding。

测试发现：无法验证；log 显示了它停在哪里。 2 分钟。

快速跳转到您 'frecently' 访问过的目录：一个带有额外功能的 z.sh 原生 Zsh 移植版本。

测试发现：无法验证；log 显示了它停在哪里。 15 分钟。

通过 TCP 连接进行 ping

测试发现：无法验证；log 显示了它停在哪里。 3 分钟。

终端中的 AI 结对编程：单个静态二进制文件，亚毫秒启动，支持任何模型

测试发现：无法验证；log 显示了它停在哪里。 7 分钟。

用于提取.zip、.rar、.bz2、.gz、.zlib、.tar、.tbz2、.tgz、.Z、.7z、.xz、.exe、.tar.bz2、.tar.gz、.tar.xz 等格式的 Bash/Zsh 函数

测试发现：无法验证；log 显示了它停在哪里。

PHP 性能工具，分析脚本的时间、内存使用和数据库查询。支持 Laravel 和 Composer，适用于 Web、Web 控制台和命令行界面。

测试发现：无法验证；log 显示了它停在哪里。 1 分钟。

用于 AI coding agents 的本地代码库智能 CLI + MCP 服务器：SQLite 代码图、28 种语言、287 个命令、246 个 MCP 工具、变更安全门、审计证据、zero API 密钥。

测试发现：无法验证；log 显示了它停在哪里。