A terminal-based AI coding assistant that provides interactive refactoring, automated reviews, and background execution via the Cursor agent.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install cursor-agent-cli
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install cursor-agent-cli using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is Cursor Agent CLI?
The Cursor Agent CLI skill integrates the advanced AI capabilities of the Cursor editor directly into the developer's terminal. By providing a dedicated agent tool, it allows for seamless interactive programming sessions where the AI can analyze, plan, and execute code changes. This skill is a core component of the Openclaw Skills ecosystem, enabling developers to bridge the gap between manual terminal commands and sophisticated AI-driven automation.
At its core, the skill utilizes two primary tools: the cursor launcher for editor integration and the agent for autonomous task execution. It supports a variety of high-performance models like GPT-4 and Claude 3.5 Sonnet, offering specialized modes for planning, asking questions, or full-scale code modification. Whether you are performing a complex migration or generating unit tests, this integration ensures your CLI remains the center of your productivity.
Cursor Agent CLI Use Cases
- Rapid code refactoring to update legacy modules or implement modern design patterns.
- Automated bug identification and resolution by providing the agent with error logs or specific file contexts.
- Security-focused code reviews of uncommitted changes using git diff pipes.
- Generating comprehensive test suites and technical documentation for entire project directories.
- Background execution of long-running development tasks using the Cloud Agent feature.
How Cursor Agent CLI Works
- The user initializes the skill by installing the Cursor CLI and authenticating through the login command.
- When a command is issued via the agent, the tool scans the local workspace to build a context of the codebase.
- The agent processes the natural language prompt against the selected LLM, determining the necessary file modifications or terminal actions.
- If in Plan mode, the agent proposes changes for review; in Agent mode, it requests approval before applying edits.
- Changes are executed locally or in a sandbox, with session history preserved in Openclaw Skills for future reference or resumption.
Cursor Agent CLI Setup
To install the Cursor CLI tools, execute the appropriate command for your operating system:
# macOS, Linux, WSL
curl https://cursor.com/install -fsS | bash
# Windows PowerShell
irm 'https://cursor.com/install?win32=true' | iex
Once installed, verify the installation and authenticate your session:
agent --version
agent login
Cursor Agent CLI Data Schema & Taxonomy
The skill organizes its configuration, cache, and session metadata within the user's home directory while utilizing environment variables for headless operations.
| Data Type | Location / Source | Purpose |
|---|---|---|
| Configuration | ~/.cursor/ |
Stores global settings, user preferences, and local cache. |
| Session History | agent ls |
Manages the metadata and identifiers for past interactive chats. |
| Worktrees | ~/.cursor/worktrees/ |
Stores temporary git worktrees used for isolated feature development. |
| Auth Token | CURSOR_API_KEY |
Environment variable used for authenticating in CI/CD or non-interactive environments. |
Cursor Agent CLI Advanced Features
- Cloud Agent Support: Offload heavy refactoring or testing tasks to background cloud workers for asynchronous execution.
- Model Context Protocol (MCP): Connect and manage MCP servers to extend the agent's capabilities with custom tools and data sources.
- Isolated Worktrees: Automatically create and manage git worktrees to keep feature development separate from the main branch.
- Force/Yolo Mode: Enable auto-approval of AI-proposed changes for trusted automated workflows and CI/CD pipelines.
- Interactive Sandbox: Control and configure the execution environment through the specialized /sandbox menu for secure code execution.
SKILL.md
Loading
Related Openclaw Skills
Consensus Send Email Guard
A persona-weighted governance system that ensures AI-generated emails meet policy and safety standards before being sent.
Consensus Persona Respawn
A persona lifecycle management tool that replaces underperforming AI evaluators with optimized successors based on historical mistake patterns.
Consensus Persona Generator
A specialized utility for generating and persisting diverse multi-agent persona sets to ensure unbiased consensus in AI workflows.
Consensus Persona Engine
A deterministic state transition layer for applying reputation updates and emitting explicit reputation delta artifacts.
Weekly Report Generator
An automated workflow that fetches team data, leverages AI for content synthesis, and generates professional Word documents for weekly reporting.
X Followings Digest Generator
A sophisticated AI automation tool that scrapes your X/Twitter following list and generates structured, categorized digests based on custom timeframes.
Featured*
