← Hub
⚡ HELP KEEP ONLINESENTINEL PIPELINE STG-1
PROMPT ENGINEERING GAUNTLET
LEVEL 1: SQL INJECTION VECTOR INGESTION ACTIVE
Our basic WAF missed a raw SQL Injection payload. Command the API engine to actively filter these logs and identify the attacker's IP.
001192.168.1.45 - - [09/Apr/2026:10:01:23 -0400] "GET /index.html HTTP/1.1" 200 1024
002192.168.1.88 - - [09/Apr/2026:10:02:11 -0400] "GET /css/styles.css HTTP/1.1" 200 3042
00310.0.0.5 - - [09/Apr/2026:10:05:01 -0400] "POST /api/healthcheck HTTP/1.1" 200 45
004192.168.1.92 - - [09/Apr/2026:10:08:33 -0400] "GET /about HTTP/1.1" 200 5120
005192.168.1.105 - - [09/Apr/2026:10:45:11 -0400] "POST /login.php?user=admin'%20OR%201%3D1-- HTTP/1.1" 200 8942
006192.168.1.45 - - [09/Apr/2026:10:18:25 -0400] "GET /login.php HTTP/1.1" 200 3102
00710.0.0.5 - - [09/Apr/2026:10:25:01 -0400] "POST /api/healthcheck HTTP/1.1" 200 45
008192.168.1.109 - - [09/Apr/2026:10:30:11 -0400] "GET /assets/logo.png HTTP/1.1" 200 89
ENGINE CONSTRAINTS (MASTER PROMPT)
Write a robust behavioral rule-set to command the AI to isolate the attacker. Warning: If it is too specific, subsequent stages will bypass this filter.
Mission Objective
Engineer a universal hunting constraint. Your single master prompt must sequentially detect raw SQL Injections, URL-encoded path traversals, and behavioral brute-force patterns across all 3 levels.
*The absolute minimum requirement is instructing the engine to find the attack. The engine will auto-extract the IP for you.*
Gauntlet Evaluation Status
⏳ TEST LAYER 1
Awaiting deployment...
🔒 TEST LAYER 2
Locked
🔒 TEST LAYER 3
Locked