Security

Security and methodology

AI Vendor Packet is built for vendor review evidence with official source links. This page explains the current security posture and where human review remains required.

Last reviewed: May 21, 2026

Control
Private app routes
Current posture
Workspace, billing, admin, findings, and report pages require an authenticated session.
Control
Session handling
Current posture
Authenticated sessions use HTTP-only cookies and server-side access checks.
Control
Rate limits
Current posture
Scanner routes and public API endpoints have basic request limits to reduce abuse.
Control
Input boundaries
Current posture
Scanner parameters and review notes are normalized before use.
Control
Audit trail
Current posture
Finding review changes are recorded with previous and next status, applicability, actor, note, and time.
Control
Secret handling
Current posture
Server-only keys should not be exposed to browser code. Client-side files are checked for secret environment names.

Methodology

Area
Source selection
How AI Vendor Packet handles it
AI Vendor Packet starts with official vendor documentation or clearly identified primary sources.
Area
Change review
How AI Vendor Packet handles it
Findings separate source text, impact hypothesis, unknowns, and suggested review actions.
Area
Uncertainty
How AI Vendor Packet handles it
If applicability is unclear, the product should say so and route the item to human review.
Area
Product boundary
How AI Vendor Packet handles it
The product supports review packets with official source links and evidence organization. It does not provide legal opinions or compliance certification.

Reference points

These public FTC resources inform the product security checklist. They do not replace a company-specific security program.

AI Vendor Packet organizes official-source review evidence and suggested review actions. It does not provide legal advice. Security controls should be reviewed before processing sensitive customer data.