Azure Ai Projects - Microsoft Foundry SDKs

This skill appears to be documentation for the Azure AI Projects Python SDK and the code examples are plausible for that purpose — but be aware of two practical issues before installing/using it: 1) Metadata mismatch: the registry metadata claims no required environment variables, but the SKILL.md expects AZURE_AI_PROJECT_ENDPOINT, AZURE_AI_MODEL_DEPLOYMENT_NAME and various connection env names. Expect to provide Azure endpoints/connections and verify which secrets are actually required. 2) Credential exposure: examples use DefaultAzureCredential, which will try local credentials (Azure CLI tokens, environment variables, managed identities). Only run this skill in environments where those credential sources are safe to use. If you have sensitive Azure credentials on the host, be cautious. 3) Data handling and tools: examples show uploading files, creating vector stores, using CodeInterpreterTool (executes Python), and calling OpenAI-compatible evals — uploaded or evaluation data may be transmitted to external services. Don’t upload sensitive files or enable code-execution tools in untrusted contexts. 4) Source and provenance: the skill lists no homepage and source is unknown. Consider obtaining the official SDK docs or a skill from a trusted publisher, and verify required env vars and intended behavior before use. If you want to proceed, confirm which environment variables you will set, review the DefaultAzureCredential auth flows you expose, and avoid uploading or evaluating any sensitive data until you’re confident about where it will be sent and who can access it.