도구

CLI and library for evaluating and red-teaming LLM apps: compare prompts and model outputs in bulk, scan for security vulnerabilities and plug into CI/CD, with evals running locally. Now part of OpenAI and still open source under MIT.

Google's official skill library and plugin marketplace: nearly 150 skills focused on Google Cloud (GKE, BigQuery, Cloud Run, IAM and more), plus 17 product plugins pairing skills with MCP for Cloud SQL, Spanner, AlloyDB, BigQuery, GCS and others.

Manages BigQuery datasets, tables and jobs with the bq and gcloud CLIs, runs SQL queries and loads data, with references for client libraries, MCP and Terraform. Google's official skill; requires a Google Cloud project.

Google's official skill library, focused on Google Cloud: 100+ skills for gcloud, Cloud Run, GKE, BigQuery, IAM, monitoring and logging, and the Well-Architected Framework, plus Ads and Analytics.

Industrial-grade open-source speech recognition toolkit from Alibaba's ModelScope, with an agent-friendly funasr command that turns audio into text, JSON or SRT subtitles in one step, with speaker diarization and timestamps. Strong on Chinese. Requires PyTorch.

AI2's tool for converting PDFs and scans to Markdown with a vision LLM, handling equations, tables, handwriting and multi-column layouts, suited to bulk digitizing papers and archives. Local runs need an NVIDIA GPU with 12GB+ VRAM; otherwise use a paid OpenAI-compatible remote endpoint.

All-in-one video translation and dubbing: chains speech recognition, subtitle translation, multi-speaker AI voiceover and audio-video muxing, with local or online models like Faster-Whisper, DeepSeek and Edge-TTS. cli.py supports headless batch runs. Needs FFmpeg.

Lets Claude Code push ML research forward unattended: literature review, topic selection, experiment deployment, cross-model review loops and paper writing through rebuttal, in 80+ skills. Codex acts as the independent reviewer by default; other model pairings work too.

Browser control from the browser-use team: an editable CDP connection lets agents drive the Chrome you are already using, writing missing helper functions as they go and keeping them for reuse. Also offers an MCP server. Allow remote debugging in Chrome on first use.

Context and harness engineering skills: context fundamentals, degradation diagnosis and compression, multi-agent architectures, memory systems, tool design, LLM-as-judge evaluation and self-improvement loops. Focused on principles and design patterns, not tied to any agent platform.

Context and harness engineering skills covering context compression and optimization, multi-agent architectures, memory systems, tool design and LLM-as-judge evaluation, activated automatically by task. Aimed at building production-grade agent systems.

Automattic's offline English grammar checker: checks spelling, grammar and word choice locally in milliseconds without sending your text anywhere. Homebrew installs harper-cli for batch-checking files (still experimental) and harper-ls for editors. English only.

Alibaba's open-source offline data sync tool: JSON job configs batch-sync between heterogeneous sources like MySQL, Oracle, PostgreSQL, Hive, HDFS and MaxCompute, run with python datax.py. Needs JDK 1.8+ and Python; download and unpack the official toolkit.

CLI version of the VideoCaptioner subtitle assistant: transcribes courses, lectures and interviews, uses an LLM to segment, proofread and translate, then burns subtitles into the video. Free ASR and Bing translation work out of the box; LLM refinement needs an API key. Python 3.10 to 3.12.

Lets agents call thousands of ready-made scrapers and automation Actors from Apify Store to pull data from social media, search engines, maps, e-commerce sites and more. Maintained by Apify; needs an Apify account (OAuth) or API token.

ModelScope's official training framework for LLMs and multimodal models: the swift command covers pretraining, fine-tuning, RL, inference, evaluation, quantization and deployment for 600+ text and 400+ multimodal models, with Megatron parallelism and a web UI. Usually needs a GPU.

Version datasets and models the way Git versions code: data lives in remote storage like S3 or GCS while versions are tracked in your Git repo. Also offers reproducible pipelines and local experiment tracking, rerunning only the steps affected by a change.

Writes precise prompts for a specific AI tool: identifies the target tool, extracts your intent, asks at most three questions if details are missing, then outputs a lean, paste-ready prompt. Covers dozens of tools including ChatGPT, Claude Code, Cursor, Midjourney and Sora.

Bundles a whole web page with its CSS, images, fonts and scripts into a single HTML file that opens offline exactly as it looked, good for archiving pages or handing them to an agent for offline analysis. It does not run JavaScript, so pre-render dynamic pages with a headless browser.

Open-source zero-shot TTS via f5-tts_infer-cli: from a few seconds of reference audio it reads Chinese or English text in the same voice, handy for lessons and videos. Needs Python, PyTorch and FFmpeg, ideally a GPU. Pretrained models are non-commercial; get consent before cloning a voice.

Turns 18 kinds of sources, including doc sites, GitHub repos, PDFs and videos, into skill packages for Claude and other platforms, or exports RAG chunks and vector store data. Optional MCP server, and it can call a local agent like Claude Code to enrich content. Needs Python 3.10+.

Open-source grammar, spelling and style checker for 20+ languages including English, German, French, Spanish and Portuguese. Proofread drafts in bulk from the CLI or run a local HTTP checking server for agents. For writers, translators and language teachers. Needs Java 17 (Homebrew includes it).

FormatJS's official localization review skill for software translators: read-only checks of UI strings and translations for ICU syntax, placeholders, terminology consistency and locale formats for dates, numbers and currency, listing issues with evidence and minimal fixes. Never edits files.

Nearly 100 skills spanning the AI research workflow: ideation and literature review, fine-tuning and post-training, distributed training, inference serving, evaluation, interpretability and paper writing, tied together by an autoresearch orchestration skill.