Skip to main content
Glama

desktop_launch

Launch a desktop application and return once its process is running. Optionally wait for a window, expose a DevTools endpoint, or start a fresh instance.

Instructions

agent-desktop launch: Launch application and return once its process is running; --activate waits for a window, --cdp verifies a DevTools endpoint

Tool call template

Replace each placeholder with a value matching the input schema. Only required parameters are included; add optional parameters as needed.

{
  "id": 1,
  "jsonrpc": "2.0",
  "method": "tools/call",
  "params": {
    "arguments": {
      "app": "<app>"
    },
    "name": "desktop_launch"
  }
}

Input Schema

NameRequiredDescriptionDefault
appYesApplication name or bundle ID
cdpNoExpose the app's DevTools protocol endpoint on this port (0 or omitted picks a free port)
cwdNoWorking directory for the launched process
envNoEnvironment variable for the launched process
argsNoCommand-line argument passed to the launched app; repeatable, order preserved. For values starting with '-', use --arg=<value>
timeoutNoUpper bound in ms for the whole launch
activateNoBring the app forward so it presents a window, and wait for one
no_attachNoStart a fresh instance instead of attaching to a running one (Windows fails if the app is already running)

Schema Changelog

Changes observed during successful MCP inspections.

  1. First observedv0.4.0

TDQS

A3.6/5.0
Behavior4/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

With no annotations, the description carries the full burden and does the most important job: it declares when the call returns (process running, not window visible) and how "--activate" and "--cdp" change that blocking behavior. It omits error and side-effect behavior, notably the default attach-to-existing-instance semantics and the Windows failure mode, which are only discoverable in the no_attach schema text.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness4/5

Is the description appropriately sized, front-loaded, and free of redundancy?

A single dense sentence with the core verb and return condition front-loaded, then the two behavior-modifying flags. The semicolon-separated flag fragments read as CLI help text, which slightly clashes with the JSON-parameter interface but wastes no words.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness3/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

For an 8-parameter mutation tool with no annotations and no output schema, the definition covers launch timing and the two most consequential flags but never says what the call returns (a process handle or ID needed to later close the app) or what happens on timeout/failure. Adequate to call, incomplete to use confidently across a launch-then-interact workflow.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters3/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema description coverage is 100%, so the baseline is 3. The description only glosses "activate" and "cdp" and does so in CLI flag notation ("--activate", "--cdp") rather than the actual JSON property names, adding no meaning beyond the schema for the six remaining parameters (app, cwd, env, args, timeout, no_attach).

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose4/5

Does the description clearly state what the tool does and how it differs from similar tools?

States a specific verb and resource ("Launch application") and adds the crucial completion semantics ("return once its process is running"), which distinguishes it from a naive fire-and-forget launcher. It does not name or contrast with any sibling such as desktop_close_app or desktop_list_apps, so an agent must infer the lifecycle relationship itself.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines3/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

The when-to-use case is strongly implied: launch an app before interacting with it, and choose --activate when a window is needed. There is no explicit statement of when not to use it, no mention of attaching vs. starting fresh as a decision point, and no pointer to any alternative tool for already-running apps.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.