gpt-4o
Pass
Audited by Gen Agent Trust Hub on Sep 22, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill is instructed to fetch documentation from the vendor's official website at
https://www.pixazo.ai/models/gpt-4o.md. This is used to provide the agent with the most up-to-date API parameters and schemas. - [INDIRECT_PROMPT_INJECTION]: The skill defines an attack surface by ingesting external data from remote sources.
- Ingestion points: The agent is instructed to fetch markdown reference files from
www.pixazo.aiand processes JSON responses fromgateway.pixazo.aiat runtime. - Boundary markers: No specific delimiters are used to separate external content from the agent's primary instructions.
- Capability inventory: The skill utilizes network tools (curl, requests, fetch) and file system access for credential storage.
- Sanitization: No explicit sanitization or instruction-filtering is applied to the retrieved content, though the sources are restricted to vendor-controlled domains.
Audit Metadata