Add Terminal / Command-Line (Shell) Tool

[mdmubarokhosin]

Issue Type: Feature Request
Priority: Medium–High
Date: August 24, 2026
📝 Description
Problem Statement
The Mobile Agent app currently lacks a built-in terminal or command-line interface (CLI) tool. This limitation prevents the agent from executing critical developer-centric tasks, such as logging into servers via SSH, building and testing projects (e.g., npm run build, pytest), directly managing Git repositories, and installing necessary packages. While the agent successfully handles browser navigation and workspace file management, the absence of CLI capabilities significantly restricts advanced development workflows.
Proposed Feature
Integrate a secure Terminal / Shell tool that empowers the agent to:

• Execute commands within a secure, sandboxed environment.
• Support standard shell utilities and language runtimes (e.g., cd, ls, cat, grep, git, npm, python).
• Accurately capture and return standard output (stdout), standard error (stderr), and exit codes.
• Maintain persistent sessions and working directories across consecutive command executions.
• Prompt for explicit user approval before executing potentially dangerous or destructive commands.
  Key Use Cases
• Remote Management: Logging into remote servers or cloud instances (ssh, scp).
• Version Control: Cloning and updating local or remote repositories (git clone, git pull).
• Dependency Management: Installing and configuring project dependencies (npm install, pip install).
• CI/CD Workflows: Building applications, running test suites, and analyzing logs.
• Infrastructure Management: Managing databases or containerized environments (e.g., Docker).
  Security Considerations
• Isolation: All commands must run inside a strictly restricted, sandboxed environment to prevent system-level vulnerabilities.
• Directory Scoping: File-system access should be limited entirely to designated workspace directories.
• Safeguards: Destructive commands (e.g., rm -rf /) or core system-level modifications must be blocked or gated behind mandatory user authorization.
• Auditability: Comprehensive logs of every executed command should be retained for security and auditing purposes.
  Expected Outcome
  By introducing a CLI tool, the agent will be able to autonomously handle end-to-end development, deployment, and server management tasks, creating a seamless experience and eliminating the need for the user to manually switch to an external terminal.
  ✅ l
• Sandboxed shell execution environment
• Support for standard shell commands
• Return standard output, error logs, and exit codes
• Persistent session and directory state retention
• User approval flow for dangerous commands
• Comprehensive audit logs