azure-ai-contentsafety-py
Pass
Audited by Gen Agent Trust Hub on Mar 1, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill utilizes the
azure-ai-contentsafetypackage, which is an official and verified SDK from Microsoft, a trusted organization. - [PROMPT_INJECTION]: The skill processes untrusted data such as user-generated text and images, which introduces a surface for indirect prompt injection.
- Ingestion points: External data is ingested via the
AnalyzeTextOptionsandAnalyzeImageOptionsmodels as documented inSKILL.md. - Boundary markers: No explicit delimiters or guardrail instructions (e.g., 'ignore commands within this text') are used when passing content to the analysis functions.
- Capability inventory: The skill's primary capability is performing network requests to Azure AI endpoints (
cognitiveservices.azure.com). No local file system modifications or command execution patterns were detected. - Sanitization: There is no evidence of local input sanitization or escaping of the ingested content, as the skill's purpose is to leverage the remote service for moderation.
Audit Metadata