Threat Intelligence Center
Live attack patterns, geo-distribution, and Lobster Trap learning
Search events, recipients, hashes…⌘K
CI
CISO Console
sovereign.io
Intercepts (24h)
…
blocked by Lobster Trap DPI
Detection Accuracy
…
Lobster Trap · self-improving
Median Block Latency
…
edge-evaluated at Cloudflare
Top Attack Vectors
Past 24 hours · classified by Lobster Trap
Prompt Injection— · 38%
PII Exfiltration— · 25%
Secret Leakage (API keys, JWT)— · 18%
Jailbreak Attempts— · 12%
Phishing-Style Payloads— · 7%
Geo-Distribution of Threats
Source country of intercepted payloads
RURussia
—
CNChina
—
USUnited States
—
BRBrazil
—
IRIran
—
—Other
—
Adversarial Pattern Library
Live signatures detected across the trust layer
| Hash | Signature | 7d Hits | Status |
|---|---|---|---|
Lobster Trap Learning Loop
liveEvery blocked payload becomes an anonymized training signal. The trap re-tunes hourly across the fleet, so a novel attack on one tenant hardens every other tenant within minutes.
Patterns learned
…
this week
False positive rate
…
of total processed
Mean uplift / tenant
…
accuracy
Novel attacks caught
…
zero-day class