BotGuard Defense.
BotGuard is an autonomous security layer designed to protect your intelligence infrastructure from malicious actors, prompt injection, and automated scraping patterns.
Injection Block
mTLS Handshake
Private Plane
01. Prompt Injection Firewall
As agents become more autonomous, the "Prompt Injection" attack vector becomes critical. BotGuard intercepts all incoming requests to the Hub and scans for adversarial patterns.
Real-time Scanning
The FirewallMiddleware inspects JSON payloads for common jailbreak strings like "ignore previous instructions" or "DAN mode".
Strict Interception
Violating requests are dropped immediately with a 403 Forbidden alert, preventing the malicious prompt from reaching your model.
02. Persistent IP Reputation
MeltyBase maintains a sovereign IP blacklist registry that tracks behavior across project boundaries. This ensures that a single malicious actor is neutralized across your entire stack.
- Dynamic Blocking: IPs that trigger the injection firewall multiple times are automatically added to the blacklist.
- Manual Governance: Enterprise administrators can manually "Master Block" problematic ranges via the MeltyBase Studio.
03. Autonomous Suppression
BotGuard doesn't just notify; it acts. The system uses a background CleanExpired worker to manage the lifecycle of blocked actors.
Bot Classification
Permit legitimate search crawlers (Googlebot, Bingbot) while aggressively suppressing unauthorized scrapers and headless browsers.
Usage Stats
Track real-time defense metrics, including TotalBlocked and TopOffenders, via the Boardroom dashboard.
Sovereign Security
Unlike third-party CDNs or firewalls, BotGuard runs natively on your MeltyBase Hub. This ensures that your security logs and actor signatures never leave your sovereign infrastructure, maintaining compliance with the most stringent data residency requirements.