도구

Anthropic's terminal coding agent: understands your codebase, edits code, runs commands and handles git workflows, with plugins, MCP and a headless mode (claude -p). Sign in with a Claude account or set an API key.

Reviews pull requests with several parallel subagents, scores each finding by confidence to filter out false positives, and can post the results as PR comments. Official Anthropic plugin; requires a logged-in GitHub CLI.

OpenAI's coding agent that runs locally, written in Rust: reads and writes code and runs commands inside an OS-level sandbox. Sign in with a ChatGPT plan or use an API key.

Open-source browser agent project. Its browser-use CLI lets Claude Code, Codex and other agents drive your logged-in local Chrome with Python snippets, or connect to a cloud browser; also ships a Python agent library. The cloud browser needs an API key.

Controls your local Chrome over CDP to navigate, click, type, take screenshots and scrape data, reusing logged-in sessions. Official browser-use skill; install the browser-use CLI with uv first.

GitHub's official CLI: manage pull requests, issues, releases and Actions workflows from the terminal. Most commands support --json output that agents can parse, and an official agent skill is available. Run gh auth login first.

Lets agents drive a real browser through the accessibility tree: navigate, click, fill forms and take screenshots, without needing a vision model. Maintained by Microsoft.

GitHub's official MCP server: lets agents read and write repositories and code, manage issues and pull requests, and check Actions workflows and code-scanning alerts. Use the hosted remote server or run it locally in Docker; needs a personal access token or OAuth.

Skills from security firm Trail of Bits: 40+ plugins for code auditing, CodeQL and Semgrep static analysis, differential review, smart-contract and cryptography audits, fuzzing and more. Also installable as a Claude Code or Codex plugin marketplace.

Cloudflare's official remote MCP servers: manage Workers and bindings such as KV, R2 and D1, and query logs and observability, builds, DNS analytics and Radar data. Authorizes with Cloudflare OAuth; some features need a paid plan.

Forces red-green-refactor: write a test and watch it fail, then write the minimum code to pass. Any code written without a verified failing test gets deleted and redone. From the Superpowers workflow.

A disciplined process for hard bugs and performance regressions: build a reliable reproduction loop, reproduce and minimize, form hypotheses, instrument to verify, then fix and add a regression test, redacting secrets in output. By Matt Pocock.

Self-improving agent from Nous Research with built-in memory and a learning loop that turns experience into skills. Works with any model provider; the hermes command has a terminal UI, and a gateway connects it to Telegram, Slack and more and runs scheduled tasks.

Drives local web apps with Python Playwright scripts to verify features, debug the UI, take screenshots and read browser logs. Includes helper scripts to start and stop dev servers. Official Anthropic skill.

Generates hook rules from a one-line description or by analyzing recent conversations. Rules are Markdown files with match conditions that block or warn on dangerous commands and unwanted patterns, and take effect without a restart. Official Anthropic plugin.

Six specialized review subagents that check comment accuracy, test coverage, silent failures, type design, code style and simplification opportunities. Run them individually or all at once. Official Anthropic plugin.

Skill pack from YC president Garry Tan: slash commands such as /office-hours, /plan-ceo-review, /review, /qa and /ship play CEO, engineering manager, designer, QA and release engineer. Needs Git and Bun; clone the repo and run ./setup as the README describes.

Local MCP server built into Browser Use: lets agents navigate, click, type and extract page content, or hand multi-step tasks to the built-in browser agent. Bring your own OpenAI or Anthropic API key.

Google's official open-source terminal agent using Gemini models with a 1M-token context. Built-in Google Search, file and shell tools, plus MCP support. Personal Google accounts get a free quota.

Engineering skill collection spanning define, plan, build, verify, review and ship, with /spec, /plan, /build, /test, /review and /ship commands, including TDD, code review and web performance audits. By Addy Osmani.

Addy Osmani's production-grade engineering skill pack covering spec, planning, implementation, testing, review and shipping, with commands like /spec, /plan, /build, /test, /review and /ship. Supports both Claude Code and Codex.

Microsoft's cross-browser end-to-end testing framework: one API drives Chromium, Firefox and WebKit, npx playwright test runs tests in parallel and isolation, with auto-waiting and trace replay for debugging failures.

High-throughput LLM inference and serving engine: vllm serve starts an OpenAI- and Anthropic-compatible server in one command. Supports 200+ model architectures from Hugging Face, quantization and distributed inference. Aimed mainly at NVIDIA, AMD and other GPUs.

Turns natural-language descriptions, Mermaid or a real code repo into interactive single-file HTML diagrams: architecture, flowcharts, sequence diagrams, data flows and state machines. Validates output automatically and exports PNG or SVG.