VOID OS // CHILD SAFETY — AUTOMATION LAYER (V2)
Mode: Low-noise alerts → early flags → human-in-the-loop
We’re not spying. We’re setting tripwires that surface risk early.
◼ ARCHITECTURE (LEAN)
4
[Device] → [Router/DNS Filter] → [Rules Engine] → [Alerts] → [Human Review]
◼ INPUT SIGNALS (WHAT WE WATCH)
New contacts (especially unknown adults)
Platform shifts (public → private apps)
Time anomalies (late-night activity)
Language cues (“don’t tell”, “secret”, “move here”)
App installs (new chat apps)
◼ RULE ENGINE (KEEP IT SIMPLE)
IF (NewContact AND PrivateMessaging) → ALERT_LOW
IF (PlatformSwitch) → ALERT_HIGH
IF (LateNightMessaging ≥ threshold) → ALERT_MED
IF (SecrecyLanguageDetected) → ALERT_CRITICAL
IF (NewChatAppInstalled) → ALERT_MED
Fewer rules, clearer signals. Avoid “alert fatigue.”
◼ IMPLEMENTATION (NO-CODE FIRST)
1) Network filter (whole-house)
Use Cloudflare Family DNS or OpenDNS
Blocks risky domains automatically
Gives a baseline log of domains accessed
2) Device controls
iOS: Screen Time
Android: Google Family Link
Turn on:
app install approvals
screen-time windows
content restrictions
3) App-level settings (high-risk apps)
Snapchat: restrict contacts, review friends
Discord: disable DMs from strangers, review servers
TikTok: restrict DMs, private account for minors
◼ ALERT DELIVERY (DON’T OVERCOMPLICATE)
Start with email or phone notifications
Trigger only on:
new contact DM
platform switch
secrecy keywords (manual review)
◼ HUMAN REVIEW LOOP (THE REAL CORE)
Daily (5–10 min):
- new contacts
- recent messages (spot-check)
- app installs
Weekly:
- usage patterns
- time-of-day shifts
◼ RESPONSE PLAYBOOK (ON ALERT)
Capture: screenshots timestamps
Stabilize: calm conversation, no blame
Contain: block contact, keep evidence
Report if needed:
National Center for Missing & Exploited Children
local law enforcement (urgent risk)
◼ GUARDRAILS (IMPORTANT)
Be transparent about monitoring. No hidden surveillance.
Aim for safety trust, not control.
Minimize data collection to what you actually review.
◼ FAILURE MODES (AVOID THESE)
Too many alerts → you ignore them
No rules → nothing surfaces
Secret monitoring → trust collapses
Waiting for “proof” → intervention comes late
◼ STATE
Visibility ↑
Secrecy ↓
Intervention Speed ↑
Risk ↓