link-based-auth
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill serves as an informational reference pattern for developers to implement authentication flows. It correctly utilizes secure cookie flags (HttpOnly, Secure, SameSite=Strict) and server-side validation logic.
- [INDIRECT_PROMPT_INJECTION]: The implementation examples describe a system that processes untrusted data from URL parameters.
- Ingestion points: The Edge Auth Function in
SKILL.mdingests atokenfrom the request query string. - Boundary markers: The code includes an explicit
validateToken(token)step before granting authentication state. - Capability inventory: The skill describes setting session cookies and performing data queries based on the resulting authentication state.
- Sanitization: The pattern relies on backend validation of the token to map it to a specific virtual identity, preventing unauthorized access via URL manipulation.
Audit Metadata