API with poorly scoped object access
Reconstruction of a typical REST API incident: an actor abused a predictable identifier and an export endpoint with incomplete authz. Not a real client; this is the report format I deliver.
How I work
01
Containment and evidence
Isolate the active surface, preserve logs, tokens, and artifacts without trampling the scene. Stop the bleeding first; explain second.
02
Timeline and entry vector
Order what happened and when: abused credential, stolen session, IDOR, poorly scoped flow. The “how they got in” has to be defensible.
03
Surface: endpoints and auth
Map authn/authz, poorly scoped objects, flow abuse, and trust chains. Which endpoint returned what, under which identity.
04
Impact and risk range
Classify exposed data (PII, tokens, records, secrets) and severity from low to critical, with a clear rationale for business and engineering.
05
Close and report
Revoke, rotate, patch controls, add alerts, and deliver a report that states what happened, what was or could be stolen, and what still needs hardening.
- Entry vector
- Valid support-role session plus access to other tenants’ resources via IDOR on resource identifiers.
- Observed methods
- Resource ID enumeration
- Abuse of read and export endpoints
- Session reuse with lateral privilege
- Endpoints involved
GET /api/v1/users/{id}GET /api/v1/accounts/{id}/documentsPOST /api/v1/export
- Data at risk
- Account PII (name, email, attached documents), export metadata, and session tokens potentially reusable until rotation.
- Risk range
- High to critical — multi-tenant exposure of personal data and bulk extraction via export.
- Response and close
- Mass revocation of sessions for the affected role, rotation of signing secrets, tenant-scoped authz hardening on all three endpoints, export rate limits, alerts on ID enumeration, and review of access logs in the incident window.
If you need an audit or post-incident response — we should talk. Connect