Agent Approve for Claude Code
Centralized approvals, guardrails, and activity monitoring for Claude Code from your phone
Agent Approve gives you centralized control over Claude Code from your phone. Approve or deny tool calls and shell commands from your iPhone or Apple Watch, with push notifications so you know when your agent needs a decision.
You get one shared set of guardrails, one activity history, and one approval workflow that works the same way across Claude Code, Cursor, Gemini CLI, Codex, OpenCode, OpenClaw, and Copilot.
Quick start
Download Agent Approve from the App Store and complete onboarding.
Run npx agentapprove on your machine and select Claude Code.
Scan the QR code to pair with your phone.
Start a Claude Code session — approval requests arrive on your phone or watch.
Install command
npx agentapproveAgent Approve config: ~/.agentapprove/
Claude Code hook config: ~/.claude/settings.json
Recommended configuration
- •Run
claude --dangerously-skip-permissionsto let Agent Approve handle all approvals (trusted environments only).
Capabilities
Good to know
Hook coverage – Claude Code hooks support PreToolUse (approval), PermissionRequest, Stop, and session lifecycle events.
Follow-up commands – When Claude stops, you can send new instructions from your phone to continue the session.
Headless visibility – For non-interactive use (e.g., claude -p "task" in a script), Agent Approve provides visibility into what the agent is doing and policy enforcement even when no terminal is visible.
Bulk command parsing – Policy evaluation parses shell commands chained with && into individual sub-commands evaluated independently. A dangerous command like rm -rf hidden as the third step in a chain is caught even when the batch starts with safe commands.
Stay up to date – Run npx agentapprove again at any time to update hooks to the latest version.
Use cases
Step away from your desk and approve Claude Code actions from your phone or Apple Watch instead of staying at the terminal.
Run claude -p in a script or loop and get alerts on your phone when it needs approval or finishes.
Block destructive shell commands (rm -rf, DROP TABLE) even when they are buried in a multi-command chain — one deny rule applies to Claude Code and all your other agents.
Send a follow-up prompt via voice from your Apple Watch when Claude finishes a task.
Questions about Claude Code
A few common questions about how Agent Approve fits alongside Claude Code.