azure-ai-translation-text-py
Warn
Audited by Gen Agent Trust Hub on Oct 24, 2026
Risk Level: MEDIUMMETADATA_POISONINGINDIRECT_PROMPT_INJECTION
Full Analysis
- [METADATA_POISONING]: The skill YAML frontmatter contains misleading information, specifically labeling the risk as 'critical' without justification from the skill's actual functionality, and using a future date ('2026-02-27') for the
date_addedfield. - [INDIRECT_PROMPT_INJECTION]: The skill processes external text for translation and linguistic analysis, creating an attack surface for indirect prompt injection.
- Ingestion points: External data is ingested through the
bodyargument in several SDK methods liketranslateandtransliterateas shown inSKILL.md. - Boundary markers: The skill does not provide instructions or examples of using delimiters or specific warnings to ignore instructions within the text being processed.
- Capability inventory: The skill's capabilities are limited to calling the Azure AI Translation network API. No local command execution or file modification tools are present.
- Sanitization: There is no mention of input sanitization, character escaping, or content filtering before data is sent to the translation service.
Audit Metadata