agentsclimarketplace

Simulang

Skill simular-ai/simulang-plugin/skills/simulang

Automate desktop tasks using Simulang, a JavaScript-based DSL for Simular Pro. Use when the user asks about desktop automation, UI scripting, controlling applications, keyboard/mouse automation, or writing Simular Pro scripts.From its SKILL.md

Install
npx -y skills add simular-ai/simulang-plugin --skill simulang

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

2 things to look at

  • no licenseNo license file was found in the repository. Code published without one is not open source by default, so using it at work is a question for whoever answers licensing questions where you are.
  • 1 stars1 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

SKILL.md

3.8 KB, as published. Nobody here has run it

Simulang (Simular Pro Action Language)

Simulang is a JavaScript-based DSL for Simular Pro that controls desktop environments through natural language-like functions for keyboard, mouse, perception, and application control.

Quick Reference

CategoryKey Functions
Applicationopen({app, url})
Keyboardtype({text, withReturn}), press({key, cmd, ctrl, shift})
Mouseclick({at, mode, spatialRelation}), scroll({direction})
PerceptionpageContent(), ask({prompt, context}), conceptsExist({concepts})
Waitwait({waitTime}), waitForConcepts({concepts})
Userrespond({message, requireConfirm})
FilesreadFile({path}), writeToFile({text, path})
Google SheetsgetGoogleSheetCellValue({cell}), setGoogleSheetCellValue({cell, value})
Background Browserbrowser.newtab(url), page.click(), page.type(), page.content(), page.ask()

Click Modes

// Default: text-based grounding
click({at: "sign in button"})

// Vision: for images/icons without text labels
click({at: "logo icon", mode: "vision"})

// Text + Screenshot: combines both methods
click({at: "Introduction in Simular Browser section", mode: "textAndScreenshot"})

// Spatial relation: disambiguate similar elements
click({at: "close button", spatialRelation: "containedIn", anchorConcept: "dialog"})

Spatial options: closest, furthest, above, below, left, right, contains, containedIn

Common Pattern: Extract & Process

function main() {
    open({url: "https://example.com"});
    wait({waitTime: 3});
    
    var content = pageContent();
    var result = ask({
        prompt: "Extract all items. Return as JSON array.",
        context: content
    });
    
    return JSON.parse(result);
}

Background Browser Control

For web automation tasks that can run in parallel without GUI focus, use the background browser API:

async function main() {
    // Open multiple pages in parallel
    const pages = await Promise.all([
        browser.newtab("https://news.google.com/"),
        browser.newtab("https://www.bbc.com/news")
    ]);

    // Wait for pages to load
    await Promise.all(pages.map(p => p.wait({ waitTime: 2 })));

    // Get content from all pages
    const contents = await Promise.all(pages.map(p => p.content()));

    // Summarize with LLM
    const summary = await pages[0].ask({
        prompt: "Summarize the key headlines",
        context: { text: contents.join("\n\n") }
    });

    console.log(summary);
}

Key Differences from Desktop Mode:

  • Use browser.newtab(url) instead of open({url})
  • All actions are async (await required)
  • Pages run independently in parallel
  • Use page.ask() with context: {text: ...} format

Best Practices

  1. Wait for loads: Use wait({waitTime: 2-3}) after navigation
  2. Be specific: Use role+value in descriptions ("sign in button" not "button")
  3. Use ask() for extraction: Pair with pageContent() for structured data
  4. Handle errors: Use respond({message, requireConfirm: true}) for confirmations
  5. Parallelize when possible: Open multiple browser tabs concurrently for faster results
  6. Minimize clicks: Use URLs with parameters to navigate directly when possible

Resources

Keep looking

Skills are one crate of 608,475. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.