claw-gatekeeper（爪守门人）

🛡️ OpenClaw Guardian

The Safety Brake for OpenClaw - Session-aware protection for risky operations

Overview

Claw-Gatekeeper is a comprehensive security control layer for OpenClaw that intercepts potentially dangerous operations and manages them according to their risk level:

Key Features

• 🛑 Smart Interception - Automatically catches risky operations
• 📅 Session-Level Approval - Approve MEDIUM/HIGH once, auto-approve similar operations for the session
• 🔒 CRITICAL Protection - Must confirm each CRITICAL operation individually
• 📝 Audit Trail - All MEDIUM+ operations logged to Operate_Audit.log
• ⚙️ Persistent Operation - Designed to be loaded as a resident skill

Risk Level Behaviors

🔴 CRITICAL (80-100) - Always Confirm

Examples:

• rm -rf / or system directory deletion
• Disk formatting (mkfs, format)
• System configuration changes
• Access to /etc/shadow, root SSH keys

Behavior:

• ❌ Must confirm EACH time individually
• ❌ No session-level approval available
• ❌ No auto-approval
• ✅ Complete audit logging

[OpenClaw] rm -rf ~/Projects/important

[Claw-Guardian] 🔴 CRITICAL RISK
⚠️  About to recursively delete directory with 1,247 files

Options:
  [y] ✅ Allow this time (will ask again next time)
  [Y] ✅✅ Always allow (add to whitelist)
  [n] ❌ Deny this time
  [N] ❌❌ Always deny (add to blacklist)

Note: Session approval NOT available for CRITICAL risks

🟠 HIGH (60-79) - Confirm or Session Approve

Examples:

• Deleting directories with many files
• Executing shell commands
• Installing skills from external sources
• Network requests to external domains

Behavior:

• ✅ Requires confirmation (first time)
• ✅ Can approve for entire session
• ✅ Session expires after 30min inactivity
• ✅ Complete audit logging

[OpenClaw] Installing skill from GitHub

[Claw-Guardian] 🟠 HIGH RISK
⚠️  Installing 'data-processor' from GitHub

Options:
  [y] ✅ Allow this time only
  [s] ✅📅 Allow for this session ⭐ RECOMMENDED
  [Y] ✅✅ Always allow (whitelist)
  [n] ❌ Deny this time
  [N] ❌❌ Always deny (blacklist)

User selects: [s]

✅ Operation approved for this session
📌 Similar HIGH risk operations will be auto-approved
⏱️  Session expires after 30 minutes of inactivity

🟡 MEDIUM (30-59) - Suggest Confirm or Session Approve

Examples:

• Creating new files
• Batch file operations (5-20 files)
• Reading sensitive directories
• Modifying configuration files

Behavior:

• ✅ Suggests confirmation
• ✅ Can approve for entire session
• ✅ Auto-allowed in loose mode
• ✅ Complete audit logging

🟢 LOW (0-29) - Auto-Allow

Examples:

• Reading files
• Listing directories
• Whitelisted operations
• Safe read-only commands

Behavior:

• ✅ Auto-allowed without confirmation
• ✅ No interruption to workflow
• ✅ Still logged if enabled

Installation

Prerequisites

Claw-Guardian is designed to be a persistent/resident skill in OpenClaw. It should be loaded at the start of every session.

Method 1: OpenClaw CLI (Recommended)

# Install the skill
openclaw skill install claw-gatekeeper

# Add to persistent skills (so it loads every session)
openclaw skill persist claw-guardian

Method 2: Manual Installation

# Copy skill package
cp claw-guardian.skill ~/.openclaw/skills/

# Add to autoload
openclaw skill load claw-guardian --persist

Method 3: Configuration File

Add to ~/.openclaw/config.json:

{
  "persistent_skills": [
    "claw-guardian"
  ]
}

Quick Start

1. Initialize Configuration

# Initialize with default settings
python3 ~/.claw-gatekeeper/scripts/policy_config.py show

2. Set Operation Mode

# Standard mode (recommended)
python3 scripts/policy_config.py mode standard

# Or strict mode for maximum security
python3 scripts/policy_config.py mode strict

3. Verify Installation

# Test risk assessment
python3 scripts/risk_engine.py file delete ~/test-file.txt

4. Check Session Status

# View current session info
python3 scripts/guardian_ui.py session

# View Operate_Audit.log
python3 scripts/session_manager.py check --lines 50

Session Management

How Session Approval Works

First Operation (MEDIUM/HIGH):
  [OpenClaw] Delete ~/temp/old-files/
  [Guardian] 🟡 MEDIUM RISK - Requires confirmation
             Options: [y] once, [s] session, [Y] always, [n] deny
  
  User: [s] Allow for this session
  
  ✅ Operation approved
  📌 Session approval granted

Similar Operations (same session):
  [OpenClaw] Delete ~/temp/more-files/
  [Guardian] 🟡 MEDIUM RISK - Session approved
             Auto-allowed (similar to previous approval)
  
  ✅ Auto-approved (no prompt)

Session Expiration

• Default timeout: 30 minutes of inactivity
• Activity: Any operation or confirmation resets the timer
• Persistence: Session state saved between OpenClaw interactions

Managing Session Approvals

# View current session
python3 scripts/guardian_ui.py session

# View active approvals
python3 scripts/session_manager.py list

# Revoke specific approvals
python3 scripts/session_manager.py revoke --type file --risk HIGH

# Clear entire session
python3 scripts/session_manager.py clear

Configuration

Operation Modes

Standard Mode (Recommended)

python3 scripts/policy_config.py mode standard

• CRITICAL: Always confirm (no session)
• HIGH: Confirm or session approve
• MEDIUM: Suggest confirm or session approve
• LOW: Auto-allow

Strict Mode

python3 scripts/policy_config.py mode strict

• All non-whitelisted operations require confirmation
• Session approval still available for MEDIUM/HIGH
• CRITICAL always per-confirmation

Loose Mode

python3 scripts/policy_config.py mode loose

• Only CRITICAL requires confirmation
• MEDIUM/HIGH auto-allowed after first session approval
• LOW always auto-allowed

Emergency Mode

python3 scripts/policy_config.py mode emergency

• Completely disables autonomous operations
• Everything requires confirmation
• Session approvals suspended

Managing Whitelists and Blacklists

# Add trusted paths
python3 scripts/policy_config.py add whitelist paths ~/Projects

# Add trusted commands
python3 scripts/policy_config.py add whitelist commands "git status"

# Add trusted skills
python3 scripts/policy_config.py add whitelist skills docx

# Block sensitive paths
python3 scripts/policy_config.py add blacklist paths ~/.ssh

Audit Logging

Operate_Audit.log

All MEDIUM and above operations are logged to ~/.claw-guardian/sessions/Operate_Audit.log:

[2026-03-12 14:30:25.123] [🟠 HIGH] [skill] allow_session: Installing data-processor@1.0.0 from github
[2026-03-12 14:31:10.456] [MEDIUM] [file] allow_session: delete ~/temp/cache (session approved)
[2026-03-12 14:32:05.789] [CRITICAL] [shell] allow_once: rm -rf ~/Projects/test (manual confirm)
[2026-03-12 14:35:15.234] [HIGH] [skill] deny_once: Installing suspicious-tool from unknown

Viewing Logs

# View recent entries
python3 scripts/session_manager.py check --lines 100

# Export to file
python3 scripts/session_manager.py check --lines 1000 > audit_export.txt

# Query with filters
python3 scripts/audit_log.py query 7 --risk HIGH --decision allow_session

Log Format

[TIMESTAMP] [RISK_LEVEL] [OPERATION_TYPE] DECISION: Details

Example:
[2026-03-12 14:30:25.123] [🟠 HIGH] [skill] allow_session: data-processor from github

Usage Examples

Example 1: File Cleanup with Session Approval

[OpenClaw] I'll clean up the temp directory

[Guardian] 🟡 MEDIUM RISK
           Operation: delete ~/temp/ (50 files)
           
           [y] once  [s] session ⭐  [Y] always  [n] deny  [N] always deny

User: [s] Allow for this session

✅ Approved for session
📌 Similar deletions will be auto-approved

[OpenClaw] Delete ~/cache/ (30 files)
[Guardian] 🟡 MEDIUM RISK - Session approved ✅
           Auto-allowed

[OpenClaw] Delete ~/.ssh/ 
[Guardian] 🔴 CRITICAL RISK
           [y] once  [Y] always  [n] deny  [N] always deny
           (Session approval NOT available)

Example 2: Skill Development Workflow

[OpenClaw] Testing my skill, need to install from local

[Guardian] 🟠 HIGH RISK
           Installing 'my-skill' from local
           
           [y] once  [s] session ⭐  [Y] always  [n] deny

User: [s] Allow for this session

✅ Session approved for skill development

[Repeated testing...]
[Guardian] Auto-approving local skill installations (session active)

Example 3: CRITICAL Operation Always Confirms

[OpenClaw] rm -rf ~/Projects/legacy-app/

[Guardian] 🔴 CRITICAL RISK
           Recursive deletion of 1,247 files including .git
           
           [y] allow ONCE  [Y] always  [n] deny  [N] always deny
           
           ⚠️  Session approval NOT available for CRITICAL

User: [y] Allow this time

✅ Approved (will ask again for next CRITICAL)

Script Reference

Session Management

# View session info
python3 scripts/guardian_ui.py session

# Check if operation is allowed (no interaction)
python3 scripts/guardian_ui.py check '{"operation_type":"file",...}'

# Interactive confirmation
python3 scripts/guardian_ui.py interactive '{"operation_type":"file",...}'

Session Manager Direct

# List session approvals
python3 scripts/session_manager.py list

# Revoke approvals
python3 scripts/session_manager.py revoke --type file --risk MEDIUM

# Clear session
python3 scripts/session_manager.py clear

# View Operate_Audit.log
python3 scripts/session_manager.py check --lines 50

Risk Assessment

# Assess file operation
python3 scripts/risk_engine.py file delete ~/test.txt

# Assess shell command
python3 scripts/risk_engine.py shell "rm -rf /tmp/*"

# Assess network request
python3 scripts/risk_engine.py network https://api.example.com POST

# Assess skill installation
python3 scripts/risk_engine.py skill my-skill github

Best Practices

For Personal Use

1. Use session approval for development work

- Approve temp file deletions for session - Approve git operations for session - Approve skill testing for session

1. Never session-approve CRITICAL risks

- Always review each CRITICAL operation - CRITICAL = potential data loss or system damage

1. Review Operate_Audit.log weekly

   python3 scripts/session_manager.py check --lines 100

For Team/Enterprise

1. Standard mode for most users
2. Strict mode for production systems
3. Regular audit log reviews
4. Document session approval policies

Troubleshooting

Session Not Persisting

Problem: Session approvals lost between interactions

Solution: Ensure skill is loaded as persistent:

openclaw skill list --persistent
# If not listed:
openclaw skill persist claw-guardian

Too Many CRITICAL Prompts

Problem: Every CRITICAL operation requires confirmation

This is by design. CRITICAL risks must always be confirmed individually. Consider:

• Whitelisting safe operations
• Reviewing why operations are marked CRITICAL
• Using less destructive alternatives

Session Timeout Too Short

Problem: Session expires during work

Solution: Adjust timeout (requires config edit):

# In ~/.claw-guardian/config.json
{
  "session_timeout": 3600  # 1 hour in seconds
}

Project Status

This is a temporary security measure.

Claw-Guardian addresses current security gaps in OpenClaw. Once OpenClaw implements comprehensive built-in safety controls, this project may be deprecated.

Claw-Guardian - Making OpenClaw Safer, One Session at a Time 🛡️