Home > Tasks > Execution > RunToolTask

RunToolTask

Executes external CLI tools and scripts within the workspace environment with real-time output streaming and manual approval gates.

Category: Execution Side-Effect: Destructive Model: GPT-4 Preferred
⚙️ Execution Config (JSON)

{
  "tool": "npm",
  "args": ["test", "--", "--watchAll=false"],
  "workingDir": "webui",
  "task_description": "Run frontend unit tests",
  "task_dependencies": []
}
→
👁️ Session UI Render
Proposed command to run
npm test -- --watchAll=false
Process Output 
PASS src/App.test.js
Test Suites: 1 passed, 1 total
Tests:       3 passed, 3 total
Snapshots:   0 total
Time:        1.2s

Workspace Artifacts

The agent maintains a persistent workspace for each task. Below is the live file structure for the RunToolTask test workspace, showing logs and generated outputs.

Execution Configuration (Runtime)

Field Type Description
tool * String The name of the tool to run. Resolved via UserSettingsManager.
args List<String> Arguments passed to the tool executable.
workingDir String Relative path of the working directory for execution.
task_description String Human-readable description of the intent.
task_dependencies List<String> IDs of tasks that must complete before this one.

Type Configuration (Static)

Field Type Description
task_type String Always RunTool.
model ApiChatModel Optional LLM override for task-specific logic.

Token Usage

Low

Dependencies

None (Standalone)