openai-sdk
Pass
Audited by Gen Agent Trust Hub on Sep 27, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill documents patterns for ingesting external data such as image URLs and user prompts which presents a surface for potential indirect instruction injection. 1. Ingestion points: Untrusted data enters the agent context through inputs in the vision and chat completion examples in SKILL.md. 2. Boundary markers: The provided code examples do not incorporate specific delimiters or instructions to ignore embedded content within the external inputs. 3. Capability inventory: The skill utilizes file system access to read local images and performs network operations via the OpenAI SDK. 4. Sanitization: The examples do not include logic for sanitizing or validating remote URLs or prompt strings.
Audit Metadata