Sentinel Configuration

Policy Configuration

Automatically control which AI tools employees can use and what data they can share:

Sentinel endpoint policy configuration dashboard

Via UI

Go to Governance → Sentinel → Policies
Click + Create Policy
Name: “Standard Employee Policy”
Scope: Select which users/departments
Approved Tools: Check which tools allowed
Blocked Tools: Check which to block
DLP Rules: Configure data protection
Click Deploy to Agents

Via Configuration File

Edit config file on deployed machine:

Windows: C:\Program Files\TruthVouch\sentinel\config.json macOS: ~/.truthvouch/sentinel/config.json Linux: /etc/truthvouch/sentinel/config.yaml

policies:
  - name: "Standard Employee"
    scope:
      users: "all"
      exclude_groups:
        - "IT Department"

    approved_tools:
      - chatgpt
      - claude
      - copilot

    blocked_tools:
      - chatgpt_free_tier
      - any_unauthorized_llm

    dlp:
      enabled: true
      block_patterns:
        - type: "ssn"
          pattern: '\d{3}-\d{2}-\d{4}'
        - type: "credit_card"
        - type: "api_key"
        - type: "password"
      action: "block"

    exceptions:
      - user: "security_researcher@company.com"
        allow: ["any_tool"]

Monitoring Configuration

Control what Sentinel monitors:

monitoring:
  # Monitor specific AI tools
  tools:
    chatgpt: true
    claude: true
    copilot: true
    gemini: true

  # What to capture
  capture:
    prompts_and_responses: true  # Log interactions
    copy_paste: true              # Monitor clipboard
    screenshots: false            # Don't capture screenshots
    browser_history: true         # Track visited sites
    file_access: true             # Monitor file reads

  # Privacy settings
  privacy:
    encrypt_logs: true
    anonymize_user_ids: false
    mask_pii_in_logs: false

Reporting Configuration

Configure how Sentinel reports back:

reporting:
  # Upload interval
  upload_interval_minutes: 60

  # What to report
  report:
    policy_violations: true
    tool_usage: true
    dlp_blocks: true
    errors_and_warnings: true

  # Compression
  compress_logs: true

  # Retry if offline
  retry_on_failure: true
  max_retries: 5

Device Configuration

Per-device settings:

device:
  name: "john-laptop-01"
  user_email: "john@company.com"
  department: "Engineering"

  # Auto-apply policies for this department
  auto_apply_department_policies: true

  # Allow offline operation?
  offline_mode:
    enabled: true
    cache_size_mb: 500
    max_hours_offline: 24

  # Auto-update
  auto_update: true
  auto_update_time: "02:00"  # 2 AM

  # Update frequency
  policy_sync_interval_minutes: 30

Logging Configuration

Control logging behavior:

logging:
  level: "info"  # debug, info, warn, error

  # Log files
  file:
    path: "/var/log/truthvouch/"
    max_size_mb: 100
    max_files: 10
    compress_old: true

  # System event log (Windows)
  windows_event_log: true

  # Syslog (Linux)
  syslog: true

Advanced Settings

Performance Tuning

performance:
  # Max CPU usage
  max_cpu_percent: 5

  # Max memory
  max_memory_mb: 256

  # Batch uploads (faster, uses more bandwidth)
  batch_size: 100

  # Queue size before dropping
  max_queue_size: 10000

Network Settings

network:
  # Proxy settings
  proxy:
    enabled: false
    host: "proxy.company.com"
    port: 8080
    username: "proxyuser"
    password: "proxypass"

  # API endpoint
  api_endpoint: "https://api.truthvouch.io"

  # Certificate validation
  verify_ssl: true

  # Timeout (seconds)
  timeout: 30

Security Settings

security:
  # Require password to view logs
  password_protect_logs: true

  # Require password to uninstall
  password_protect_uninstall: true

  # Prevent disabling
  prevent_disable: true

  # Tamper detection
  tamper_detection: true
  alert_on_tamper: true

Group Policy (Windows)

Deploy standardized config to all machines:

Group Policy Editor (gpedit.msc)

Location:
  Computer Configuration > Administrative Templates >
  TruthVouch > Sentinel Agent

Available Settings:
  - Require password to disable: [X] Enabled
  - Policy update interval: 30 minutes
  - Report upload interval: 60 minutes
  - Block external tools: [X] Enabled

Policy Sync

Policies automatically sync from TruthVouch:

Agent checks every 30 minutes (configurable)
  ↓
Downloads latest policies
  ↓
Applies immediately (no restart needed)
  ↓
Offline cache prevents policy drift

To force immediate sync:

# Windows
sentinel-agent --sync-policies

# macOS/Linux
sentinel-agent --sync-policies

Verification

Verify configuration applied:

# Windows
C:\Program Files\TruthVouch\sentinel.exe --verify-config

# macOS/Linux
sentinel-agent --verify-config

Output:

Configuration validation:
✓ Config syntax: OK
✓ Policies: 3 loaded
✓ DLP rules: 12 active
✓ Network: Connected
✓ Encryption: Enabled
Status: Ready