oidc-hosted-page
Pass
Audited by Gen Agent Trust Hub on Mar 1, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [SAFE]: The skill utilizes non-sensitive placeholders such as
<my_client_id>andYOUR_CLIENT_SECRETfor credentials, correctly instructing users to handle real secrets securely. - [SAFE]: Network references are limited to official vendor domains (e.g.,
auth.ssojet.com), which are legitimate endpoints for the described OIDC functionality. - [PROMPT_INJECTION]: Identified an indirect prompt injection surface in
PROMPTS.mdwhere user-provided values are interpolated into agent instructions. - Ingestion points: Placeholders like
{framework}and{my_api_endpoint}inPROMPTS.mdallow for external data ingestion. - Boundary markers: Missing explicit delimiters to separate user input from instructional text.
- Capability inventory: Instructions involve file system modification and package installation (
npm install). - Sanitization: No validation or escaping is applied to the placeholder content.
Audit Metadata