Cursor cli
Claude Code, Codex, OpenClaw and Ottili AI Skills - powered by Ottili ONE
npx -y skills add Ottili-ONE/skills --skill cursor-cliAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
One thing to look at
- 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.
What its author says it does
Copied from the file, not written here
Use when automating or invoking Cursor CLI (agent/cursor-agent) for non-interactive tasks, CI pipelines, or multi-agent delegation.
The file declares its own license as MIT. That is the author’s claim about this one file, and it is not the same thing as the license GitHub reports for the repository, which is listed with the other numbers below.
SKILL.md
3.9 KB, as published. Nobody here has run it
Cursor CLI Agent Skill
Guide for Cursor CLI (agent / cursor-agent) including installation, authentication, non-interactive mode, and automation patterns.
Installation
Standard (macOS, Linux, Windows WSL)
curl https://cursor.com/install -fsS | bash
Homebrew (macOS)
brew install --cask cursor-cli
After installation
Add to PATH if needed:
export PATH="$HOME/.local/bin:$PATH"
Verify:
agent --version
# or
cursor-agent --version
Authentication
Browser login:
agent login
Or API key:
export CURSOR_API_KEY=your_api_key_here
Update
agent update
# or
agent upgrade
Commands
Interactive
agent
agent "Add error handling to this API"
Backward compatible: cursor-agent works the same way.
Model selection
agent models
agent --model gpt-5
During session: /models
Sessions
- List:
agent ls - Resume last:
agent resume - Specific:
agent --resume="[chat-id]"
Context (files/folders)
@filename.ts
@src/components/
Slash commands (interactive)
/models– switch model/compress– summarize conversation/rules– create or edit rules/commands– custom commands/mcp enable [server-name]//mcp disable [server-name]
Non-interactive / CI mode
For scripts and automation, prefer print mode:
agent -p 'Run tests and report coverage'
agent -p 'Refactor this file to use async/await'
Output formats:
agent -p 'Analyze code' --output-format text
agent -p 'Find bugs' --output-format json
agent -p 'Run tests' --output-format stream-json --stream-partial-output
Force edits without confirmation:
agent -p 'Fix all linting errors' --force
Media input:
agent -p 'Analyze this screenshot: screenshot.png'
Automation from other agents
When calling Cursor CLI from another agent or script, use print mode (-p / --print) with --output-format text or json. Print mode does not require an interactive TTY.
Example for delegation:
cursor-agent --model auto -p "Review auth middleware for security issues" --output-format text
For interactive TUI workflows in headless environments, use tmux to provide a PTY:
tmux new-session -d -s cursor
tmux send-keys -t cursor "cd /path/to/project" Enter
tmux send-keys -t cursor "agent 'Your task here'" Enter
sleep 60
tmux capture-pane -t cursor -p -S -100
Rules and configuration
The agent loads automatically:
.cursor/rulesAGENTS.mdCLAUDE.md
Use /rules to manage rules inside the CLI.
MCP integration
MCP servers come from mcp.json. Enable or disable during a session:
/mcp enable server-name
/mcp disable server-name
Common workflows
Code review
agent -p 'Review the changes in the current branch against main. Focus on security and performance.'
Refactoring
agent -p 'Refactor src/utils.ts to reduce complexity and improve type safety.'
Git commit message
agent -p 'Generate a commit message for the staged changes adhering to conventional commits.'
CI/CD
export CURSOR_API_KEY=$CURSOR_API_KEY
agent -p 'Audit this codebase for security vulnerabilities' --output-format json --force
Multi-agent delegation
When used with work-untill-i-die, Cursor is typically the grunt role for fast, cheap iteration:
cursor-agent -p "<self-contained brief>" --model auto --output-format json
Add --force only when the delegate must modify files. Default to read-only delegation when possible.