- name
- web-auth-mapping
- description
- Authentication surface — login endpoints, JWT/OAuth/SAML/SSO/API-key mechanism identification.
- allowed-tools
- Bash Read Write
- metadata
- {"subdomain":"reconnaissance","when_to_use":"auth surface, login endpoints, JWT, OAuth, SAML, SSO, API key, authentication mechanism","tags":"auth-mapping","mitre_attack":"T1589.001"}
# Authentication Surface Mapping
Map every authentication mechanism the target exposes. Exploit downstream needs to know exactly how to hold a session (cookie, JWT, API key) and where to attack auth (token theft, race-condition on bcrypt, SSO redirect chain).
## 1. Login Endpoint Discovery
```bash
# Common auth paths
for path in login signin auth authenticate oauth/authorize \
api/auth api/login admin/login wp-login.php; do
code=$(curl -s -o /dev/null -w "%{http_code}" "https://<target>/$path")
[ "$code" != "404" ] && echo "$code https://<target>/$path"
done
```
## 2. Auth Mechanism Identification
| Mechanism | Signal |
|-----------|--------|
| Cookie-based | `Set-Cookie` headers after login (often `session=`, `JSESSIONID=`, `PHPSESSID=`) |
| JWT | `Authorization: Bearer eyJ...` patterns; three base64 segments separated by `.` |
| OAuth 2.0 | `/oauth/authorize`, `/oauth/token` endpoints; `state`/`code`/`redirect_uri` params |
| API Key | `X-API-Key` header accepted; `Authorization: ApiKey <token>` |
| SAML/SSO | Redirects to IdP (Okta, Azure AD, Auth0); `SAMLRequest` form param |
| Session row + slow KDF | Login latency >50ms on wrong password (bcrypt/Argon2) — flag for race-condition recon |
## 3. Output for Handoff
Record per mechanism: endpoint, success/fail signals (status code, cookie set, JSON shape), credentials used, observed latency. This feeds the **Required session state** line in the recon → exploit handoff.
GitHub에서 보기