CHAPTER 02 · ATTACK · CATCH WHAT REVIEW CAN'T SEE
Think like the intruder. Fix like the owner.
AI agents map your app, form attack hypotheses, prove what's exploitable on staging, and hand you the fix. Reports are SOC 2 and ISO ready.
White-box assessment of checkout-api
FEL-142 · Cross-tenant order read
FEL-143 · JWT alg=none impersonation
11 hypotheses disproved, with evidence
0
agents per assessment
0
targets: web, API, Android, MCP, LLM
0
false positives reported as exploits
Daily
attack-surface monitoring
01 TARGET · 02 ACCESS · 03 REVIEW & LAUNCH
Point it at staging. That's the setup.
Feluda pre-fills the target from your repo, verifies ownership through your deployment, and respects the paths you never want touched.
What should we attack?
We pre-filled this from puku/checkout-api. Change anything that looks off.
Target type
Base URL
Verified automatically via your GitHub deployment. Use staging — some tests create or delete data.
Code context (recommended)
Never touch (optional)
ESTIMATE
~4.0K requests
of 12.4K left this month · paced across 2 × 5h windows · ~3h
Endpoints mapped42
Modewhite-box
Excluded paths2
Paced to stay inside your 5h and weekly windows. Pauses safely if you hit a limit.
AI CODE ANALYSIS
A security audit of the code itself.
Pick repos and docs. See the exact request cost before you start. Results land in Pull Requests and Slack.
CODE SECURITY AUDIT · 2 REPOS
Ready when you are
total4.2K req
ATTACK SURFACE · MONITORED DAILY
Everything you expose to the internet.
Subdomain takeovers, leaked credentials, open ports and spoofable mail, each with the one action that fixes it.
YOUR NEXT CASE