understudy-aiOfficial·48 Agent Skills Included

understudy

Operates your computer, learns tasks, automates daily workflows

Operates your entire computer from a single instruction, controlling apps, browsers, shell, and messaging channels like Telegram, Slack, and WhatsApp. Learns tasks by watching you demonstrate them once, then replays them reliably even after interfaces change. Eliminates repetitive clicking, manual file handling, and multi-step routines by turning them into reusable skills and multi-stage automated pipelines.

455|34

↗

npx skills add understudy-ai/understudy --all -g -y

All Skills in This Repository (48)

Pure Emerald Level Indicators

📦 In Repo

understudy-ai

taught-person-photo-cutout-bc88ec

Remove image backgrounds via Google Chrome and Pixelmator Pro, then send to Telegram.

nano-pdf

Edit specific PDF pages using natural-language instructions via the nano-pdf CLI.

himalaya

Manage emails via IMAP/SMTP from the terminal with multiple accounts.

bear-notes

Manages Bear notes on macOS via the grizzly CLI tool.

peekaboo

Automate macOS UI interactions and system tasks via command-line interface.

model-usage

Summarize local AI model costs per model using the CodexBar CLI.

blogwatcher

Monitor blogs and RSS/Atom feeds for new articles using the blogwatcher CLI.

discord

Send, reply, edit, react to, and delete Discord messages with attachments.

coding-agent

Delegate coding tasks to AI agents like Codex or Claude Code.

openhue

Control Philips Hue lights and scenes via the OpenHue CLI tool.

gemini

Interact with Gemini AI models via command line for Q&A, summarization, and text generation.

gifgrep

Search Tenor and Giphy GIFs via CLI or TUI to preview, download, and extract frames.

Official

Intermediate

Frequently Asked Questions

FAQPage Schema

How to install Understudy?▼

Run `npx skills add understudy-ai/understudy --all -g -y` in your terminal to install all skills in this suite globally.

What can Understudy automate on my computer?▼

It controls desktop apps, browsers, and the shell to handle tasks like converting files, sending messages, editing images, and managing notes or reminders. You can also teach it new tasks by demonstrating them once.

How does teaching a task by demonstration work?▼

You perform a task once while Understudy watches, and it extracts the intent rather than memorizing clicks. The resulting skill survives UI redesigns and can be replayed later with a plain-English request.

Can I control Understudy remotely from my phone?▼

Yes. It connects to messaging apps like Telegram, Discord, Slack, WhatsApp, Signal, and iMessage, so you can send instructions from your phone and it executes them on your desktop.

Which AI models does Understudy support?▼

It works with your own API keys for providers like OpenAI, Anthropic Claude, and Google Gemini. The runtime is open source under the MIT license, so you only pay for model usage.