A specialized skill for managing OpenClaw approval permissions, allowing users to toggle between secure allowlists and open execution modes.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install permission-manager
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 permission-manager 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 Permission Manager?
Permission Manager is a critical security and configuration tool designed for the OpenClaw ecosystem. It empowers developers and users to dynamically adjust the execution safety levels of their AI agents, ensuring a perfect balance between airtight security and operational velocity. By leveraging this tool among your Openclaw Skills, you can prevent unauthorized command execution while maintaining the flexibility needed for rapid development.
The skill provides a natural language interface to modify internal security protocols. Whether you need strict oversight for production environments or frictionless automation for local testing, Permission Manager acts as the central hub for policy enforcement. It bridges the gap between complex configuration files and intuitive user intent, making high-level security management accessible to every developer.
Permission Manager Use Cases
- Switching to high-security allowlist modes for production environments to prevent unverified command execution.
- Enabling full permission mode during development and debugging to streamline automated testing workflows.
- Granting temporary 'No-Approval' status for trusted local scripts while ensuring human-in-the-loop oversight is just a command away.
- Auditing current security settings to verify compliance with organizational safety standards.
How Permission Manager Works
- The skill parses natural language input to identify the user's intended security state (e.g., allowlist, full, or no-approval).
- For high-risk transitions like 'No-Approval', the skill triggers a mandatory risk disclosure and waits for explicit human confirmation.
- Upon confirmation, it executes systemic configuration updates using the openclaw config set command structure.
- The skill validates the change by reading the updated configuration file and informs the user if a Gateway restart is required.
- Optionally, it records the user's preference in long-term memory for contextual awareness in future sessions.
Permission Manager Setup
To enable the Permission Manager within your environment, ensure you have the OpenClaw CLI installed and run the following configuration commands:
# Set the execution security level
openclaw config set tools.exec.security allowlist
# Define the approval policy
openclaw config set tools.exec.ask on-miss
Ensure your gateway has read/write access to ~/.openclaw/openclaw.json to persist these changes across sessions.
Permission Manager Data Schema & Taxonomy
The skill manages two primary configuration files to maintain security states and audit trails:
| File Path | Description | Key Fields |
|---|---|---|
~/.openclaw/openclaw.json |
Global system configuration and security policies. | tools.exec.security, tools.exec.ask |
~/.openclaw/exec-approvals.json |
A historical log of specifically approved commands and execution metadata. | command, timestamp, status |
Data is structured in a nested JSON format, allowing for granular control over individual tool execution policies.
Permission Manager Advanced Features
- Risk Disclosure Workflow: Automatically forces a confirmation step before entering dangerous low-security modes.
- Multi-Mode Support: Seamlessly switch between Default (Allowlist), Full Permission, Convenience (No-Approval), and Strict modes.
- Contextual Recovery: Quickly restore safe default settings with a single command to reset the environment after testing.
- Session Isolation: Changes can be applied to the current session immediately, with options for permanent persistence.
SKILL.md
Loading
Related Openclaw Skills
Increment Code Unit Test Creator
An AI-powered Java testing specialist that generates high-coverage JUnit 5 and Mockito tests specifically for incremental code changes identified via Git.
Alipay AI Product Intro
A specialized skill for retrieving premium product information regarding Alipay AI Agent Wallets and AI payment solutions via the A2M protocol.
Luckin Coffee Order MCP Skill
A comprehensive Model Context Protocol service for Luckin Coffee that enables automated ordering, store discovery, and account management.
Skill Auditor Plus
A professional-grade auditing tool designed to secure and optimize AI agent skills by scanning for vulnerabilities, credential leaks, and performance bottlenecks.
clawtip-weather
clawtip-weather is a professional AI agent skill that provides expert-level daily weather reports via a secure, payment-verified workflow.
AI Company CEO
A comprehensive executive management skill designed to lead all-AI organizations through structured Hub-and-Spoke governance and multi-agent orchestration.
Featured*
