agentleFS
Sign inStart free
matheus-rib1108matheus-rib1108/meu-kit-claude-code/skills/produtividade-e-pc/pc-control/SKILL.md

pc-control

Skill0 starsChanged 27 days agoMITView on GitHub

Gets Claude controlling the user's own computer and browser — opening programs (Obsidian, Notepad, VS Code…), using Chrome while logged in, creating accounts on sites, signing up for apps — with the least effort from the user. Use when the user asks Claude to "mexer no meu PC", "abrir o programa", "instalar a extensão", "criar uma conta pra mim", "entrar no site com meu login", or when a task needs their logged-in browser or a desktop app.

2 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

3 rules. About 0 are common, 3 are distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈3

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Prefer "Continuar com o Google" (no new password).

    line 23
  2. Distinctive:

    If the user pastes a password anyway, tell them to change it afterwards.

    line 26
  3. Distinctive:

    Before accepting terms, tell the user what they agree to (paid plan, trial that auto-charges, data use).

    line 29

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

PC control

Where Claude is running decides what's possible

  • Cloud session (claude.ai/code, the phone app's Code tab): runs in a remote container. It cannot touch the user's computer. Say so in one line and give the setup below.
  • Claude desktop app with computer use on: can see the screen, click, type, open any app.
  • Claude Code on the user's computer: terminal + files; with Claude in Chrome (claude --chrome) or the browser-use / chrome-devtools-mcp plugins it drives their Chrome.

One-time setup (the only thing the user must do)

  1. Install the Claude desktop app and turn on computer use (Settings). From then on Claude opens programs and does tasks itself, asking permission per app.
  2. Optional, for websites with their login: the Claude in Chrome extension from the Chrome Web Store (published by Anthropic). Chrome always asks the user to confirm an extension install — that confirmation is the user's one click; never try to skip it.

Creating accounts for the user

  1. Prefer "Continuar com o Google" (no new password). Otherwise ask which email to use.
  2. Passwords never go in the chat. Generate a strong one in the browser's password manager (Google Password Manager or Bitwarden) during signup so it's saved there. If the user pastes a password anyway, tell them to change it afterwards.
  3. Fill the form in the user's browser (computer use / Claude in Chrome). The user handles CAPTCHA, the email/SMS verification code and 2FA — those exist to prove a human is there.
  4. Before accepting terms, tell the user what they agree to (paid plan, trial that auto-charges, data use). Never enter card details without an explicit OK for that charge.
  5. Turn on 2FA for important accounts (GitHub, Vercel, Google, hosting, payments).

No account needed at all (use these first)

  • Vercel: skill deploy-to-vercel publishes without an account and gives a claim link.
  • Neon database: Claimable Neon / neon.new (skill neon).
  • GitHub: the GitHub tools in Claude Code, or gh CLI after one gh auth login.

Opening programs (computer use)

Open via Start menu/Spotlight search, wait for the window, then act. For Obsidian prefer the skill obsidian-cli (faster and more reliable than clicking). Close what you opened when done. Never touch banking or password-manager apps unless the user explicitly asks.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

Check your setup, free