Overview
The challenge provides a predict endpoint that compares inputs with a hidden target sequence based on semantic similarity.
Reconnaissance
Fuzzing time-related phrases yields score boosts. Testing the phrase time wheel reaches a 100% semantic score of 1.0.
Exploitation Strategy
However, a secondary temporal check is active on the timestamp parameter. Fuzzing timestamps against the CTF start date (2026-03-28) triggers the flag response.
Solution Code
import requests
r = requests.post("http://138.199.163.92:10675/predict", json={
"text": "time wheel",
"timestamp": "2026-03-28"
})
print(r.json())Flag
Kaal{Tareekh_pe_tareekh}