Public Repository - Security hardening and monitoring tools for Claude Code CLI environments

⚠️ SECURITY NOTICE: This repository contains configuration TEMPLATES. Never commit actual security-config.json or storage-rules.json files with your real system paths, credentials, or sensitive information. Always customize configurations locally and keep them in .gitignore.

This toolkit provides comprehensive security controls, monitoring, and hardening measures for Claude Code CLI installations. It addresses identified vulnerabilities and implements defense-in-depth strategies to protect against unauthorized operations, data exfiltration, and behavioral manipulation.

• Protect user systems from unauthorized Claude Code operations
• Implement access controls for file system, network, and command execution
• Monitor and audit all Claude Code activities
• Enforce behavioral constraints for transparent, accountable operation
• Provide corrective actions for identified security gaps

• ✅ Comprehensive PreToolUse security hooks
• ✅ Storage access policy enforcement
• ✅ Command injection prevention
• ✅ Network activity monitoring
• ✅ Session audit logging
• ✅ File integrity monitoring
• ✅ Rate limiting and anomaly detection
• ✅ Symbolic link resolution
• ✅ Behavioral constraint enforcement

# Clone this repository (after pushing to GitHub)
git clone https://github.com/hah23255/claude-code-security-toolkit.git
cd claude-code-security-toolkit

# Run installation script
bash scripts/install.sh

# Copy and customize configuration templates
cp config/templates/security-config.json ~/.claude/security-config.json
cp config/templates/storage-rules.json ~/.claude/storage-rules.json

# Edit with your specific paths and policies
vi ~/.claude/security-config.json

# Deploy enhanced security hooks
bash scripts/deploy-hooks.sh

# Test hook enforcement
bash scripts/test-hooks.sh

claude-code-security-toolkit/
├── README.md                          # This file
├── SECURITY.md                        # Security policy and reporting
├── LICENSE                            # License information
├── .gitignore                         # Git ignore patterns
│
├── docs/                              # Documentation
│   ├── reports/                       # Security assessment reports
│   │   ├── integrity-assessment.md   # Initial integrity assessment
│   │   ├── corrected-assessment.md   # Corrected user perspective
│   │   └── gap-analysis.md           # Security gap analysis
│   ├── guides/                        # Implementation guides
│   │   ├── installation.md           # Installation instructions
│   │   ├── configuration.md          # Configuration guide
│   │   ├── hardening.md              # System hardening guide
│   │   └── troubleshooting.md        # Troubleshooting guide
│   └── references/                    # Reference documentation
│       ├── hook-api.md               # Hook development reference
│       ├── gap-details.md            # Detailed gap descriptions
│       └── attack-scenarios.md       # Attack scenario documentation
│
├── hooks/                             # Security hook implementations
│   ├── templates/                     # Original hook templates
│   │   ├── storage_policy_enforcer.py
│   │   ├── security_guard.py
│   │   ├── session_start.py
│   │   └── session_end.py
│   ├── enhanced/                      # Enhanced hardened hooks
│   │   ├── comprehensive_pretooluse.py
│   │   ├── enhanced_storage_policy.py
│   │   ├── enhanced_audit_logger.py
│   │   └── file_integrity_monitor.py
│   └── utils/                         # Shared hook utilities
│       ├── patterns.py               # Security pattern definitions
│       ├── logger.py                 # Logging utilities
│       └── validators.py             # Validation functions
│
├── config/                            # Configuration files
│   └── templates/                     # Configuration templates
│       ├── security-config.json      # Main security configuration
│       ├── storage-rules.json        # Storage access rules
│       ├── settings.json             # Claude settings template
│       └── behavioral-constraints.md # Output style constraints
│
├── tests/                             # Test scripts
│   ├── test-hooks.sh                 # Hook functionality tests
│   ├── test-bypass-attempts.sh       # Security bypass tests
│   ├── test-file-access.sh           # File access policy tests
│   └── test-command-blocking.sh      # Command blocking tests
│
├── examples/                          # Example configurations
│   ├── minimal-config.json           # Minimal security config
│   ├── strict-config.json            # Strict security config
│   └── custom-rules.json             # Custom rule examples
│
└── scripts/                           # Utility scripts
    ├── install.sh                    # Installation script
    ├── deploy-hooks.sh               # Hook deployment script
    ├── backup-config.sh              # Configuration backup
    ├── restore-config.sh             # Configuration restore
    ├── audit-analysis.sh             # Audit log analysis
    └── integrity-check.sh            # System integrity check

Purpose: Prevent unauthorized file system access

Features:

• Whitelist/blacklist path enforcement
• Symbolic link resolution
• Real-path validation
• Critical file protection

Coverage: Write, Edit, NotebookEdit, Read operations

Purpose: Prevent command injection and malicious shell operations

Features:

• Dangerous pattern detection
• Command whitelist validation
• Argument sanitization
• Execution logging

Coverage: Bash tool operations

Purpose: Prevent data exfiltration and unauthorized network access

Features:

• Domain whitelist enforcement
• URL validation
• Network operation logging
• Suspicious pattern detection

Coverage: WebFetch, WebSearch operations

Purpose: Detect reconnaissance and information gathering attempts

Features:

• Sensitive pattern flagging
• Search operation logging
• Pattern-based alerting

Coverage: Glob, Grep operations

Purpose: Track and control sub-agent creation

Features:

• Agent spawn logging
• Task prompt analysis
• Suspicious task detection

Coverage: Task tool operations

Purpose: Enforce transparent, clear communication

Features:

• Output style enforcement
• Ruthless clarity standards
• Code quality requirements
• Deletion principle

Coverage: All Claude Code communications

Purpose: Maintain comprehensive activity trail

Features:

• Full operation logging
• Decision tracking (allow/block)
• Timestamp and context capture
• Forensic trail creation

Coverage: All monitored operations

Purpose: Detect and prevent automated/scripted abuse

Features:

• Per-tool rate limits
• Time-window based throttling
• Anomaly detection
• Burst prevention

Coverage: All tool operations

This toolkit addresses 15 identified security gaps in standard Claude Code installations:

• ✅ GAP 1: Read tool monitoring added
• ✅ GAP 2: Bash command monitoring implemented
• ✅ GAP 4: Fail-closed storage policy with required config
• ✅ GAP 5: Security scanner properly integrated
• ✅ GAP 12: Symbolic link resolution implemented

• ✅ GAP 3: Glob/Grep enumeration monitoring
• ✅ GAP 6: File content validation
• ✅ GAP 7: Network activity monitoring
• ✅ GAP 8: Detailed audit logging
• ✅ GAP 10: Task/agent monitoring
• ✅ GAP 13: Enhanced Edit operation controls
• ✅ GAP 15: Configuration file protection

• ✅ GAP 9: Rate limiting implemented
• ✅ GAP 11: File integrity monitoring
• ✅ GAP 14: Hook enforcement validation

See docs/references/gap-details.md for detailed information.

File: ~/.claude/security-config.json

{
  "version": "1.0",
  "fail_closed": true,
  "restricted_paths": ["~/.ssh", "~/.gnupg", "~/.claude/hooks"],
  "allowed_paths": ["/storage/emulated/0/Enterprise"],
  "dangerous_bash_patterns": ["rm\\s+-rf", "dd\\s+if=", "curl.*-d\\s+@"],
  "allowed_domains": ["docs.claude.com", "github.com"],
  "rate_limits": {"Read": 100, "Bash": 20},
  "monitoring": {
    "log_all_operations": true,
    "alert_on_suspicious": true,
    "block_on_critical": true
  }
}

File: ~/.claude/storage-rules.json

{
  "rules": {
    "RULE_01": {
      "authorized": {
        "primary": [
          {"path": "/storage/emulated/0/Enterprise", "bridge": "~/storage/Enterprise"}
        ]
      },
      "denied": {
        "paths": ["/data/data/com.termux/files/home/*"],
        "reason": "Internal storage access restricted",
        "exceptions": ["~/.claude/sessions/**", "~/.claude/file-history/**"]
      }
    }
  }
}

See docs/guides/configuration.md for complete options.

# Deploy all enhanced security hooks
cd /storage/emulated/0/Enterprise/projects/claude-code-security-toolkit
bash scripts/deploy-hooks.sh

# Run comprehensive security tests
bash tests/test-hooks.sh

# Test specific bypass attempts
bash tests/test-bypass-attempts.sh

# View recent audit entries
tail -50 ~/.claude/audit.jsonl | jq .

# Analyze blocked operations
bash scripts/audit-analysis.sh --blocked

# Generate security report
bash scripts/audit-analysis.sh --report

# Verify hook integrity
bash scripts/integrity-check.sh

# Check for unauthorized modifications
bash scripts/integrity-check.sh --verify-all

1. Review Audit Logs: Check ~/.claude/audit.jsonl weekly
2. Update Patterns: Keep hooks/utils/patterns.py current
3. Test Enforcement: Run tests after any changes
4. Backup Configuration: Use scripts/backup-config.sh

# Backup current configuration
bash scripts/backup-config.sh

# Update hooks
git pull origin main
bash scripts/deploy-hooks.sh

# Test updated hooks
bash tests/test-hooks.sh

See docs/guides/troubleshooting.md for common issues and solutions.

1. Integrity Assessment - Initial system analysis
2. Corrected Assessment - User perspective correction
3. Security Gap Analysis - Detailed vulnerability assessment

All reports available in docs/reports/.

Three documented attack chains:

1. Credential Exfiltration
2. Persistent Backdoor Installation
3. Mass Data Enumeration

See docs/references/attack-scenarios.md.

1. Create hook in hooks/enhanced/
2. Add to config/templates/settings.json
3. Test with tests/test-hooks.sh
4. Document in docs/guides/

See docs/references/hook-api.md for:

• Hook input format
• Exit code meanings
• Logging standards
• Error handling

If you discover security vulnerabilities in this toolkit:

1. Do NOT open public issues
2. Contact repository owner directly
3. Provide detailed reproduction steps
4. Allow time for patching before disclosure

See SECURITY.md for full policy.

This toolkit is provided for personal use in protecting user-owned systems. See LICENSE for details.

• Installation: docs/guides/installation.md
• Configuration: docs/guides/configuration.md
• Hardening: docs/guides/hardening.md

For issues, questions, or improvements, open an issue in this repository.

This toolkit was developed in response to observed gaps in Claude Code CLI security controls. It represents user-implemented protective measures to ensure transparent, accountable, and secure operation.

Version: 1.0.0 Last Updated: 2025-10-29