Index of every page on this site: llms.txt
All skills
microsoft avatar

/azure-ai

@0d60658
by microsoftmicrosoft/skills3.2k stars
373
Run it yourself

Build and troubleshoot Azure AI Search, Speech, Document Intelligence, Content Understanding, Translator, Content Safety, and OpenAI apps. WHEN: AI Search, vector search, hybrid search, semantic search, indexer, speech-to-text, text-to-speech, fast or batch transcription, Speech SDK logs, session ID, result ID, OCR, document translation, glossary.

Use this Skill: https://skilld.dev/gh/microsoft/skills/azure-ai

Nothing lands on disk. Nothing to clean up.

Fork this Skill

Edit a local copy. It keeps the author and licence.

referencesfoundry-toolsdata-handling.md

≈1.2k tokens on demand. Your agent reads this file only when SKILL.md points to it.

Data handling for Document Intelligence and Content Understanding

For "where does my document go", "how long is it kept", "can I delete it", "does bring-your-own storage cover it" and data-residency questions. Quote the linked Learn pages to security reviewers; where Learn pages disagree, show both statements instead of choosing one.

Topic Document Intelligence (v4.0, 2024-11-30) Content Understanding (GA 2025-11-01)
Where it is processed / stored In the region of the resource Data at rest stays in the resource region; generative steps run on your model deployment, so its deployment type (global, data zone, regional) decides where that processing can happen
Analyze input and result retention Deleted after 24 hours. Learn conflict: "24 hours after an analysis operation completes" vs "from the time that you submit" Input, intermediate data and results kept "for up to 24 hours", then deleted. An older wording (inputs deleted after processing) still appears in search results
Delete early Delete Analyze Result on the result ID from Operation-Location. Learn conflict: "permanently deletes the submitted input data and analyze result" vs REST "mark … for deletion" DELETE {endpoint}/contentunderstanding/analyzerResults/{resultId} returns 204. Needs the analyzerresults/delete data action: CU Reader lacks it; CU Contributor, Cognitive Services User or Foundry User have it
Avoid service-side temporary storage Not documented Synchronous analyzeInline / analyzeBinaryInline process in memory; preview only and only for prebuilt-read and prebuilt-layout
Processing location control Not documented processingLocation. Learn conflict: data-privacy page limits it to prebuilt-read/prebuilt-layout; REST reference exposes it on every analyze call (default global)
Training / labeled data Your Blob container (SAS or managed identity) CU Studio projects keep samples and labels in your storage through a Foundry storage connection. GA analyzers keep the labeled data and use it at analysis time; 2026-06-01-preview distills it into the analyzer
Customer-managed keys (CMK) Covers models and results for resources created after 07/31/2023; Key Vault in the same region and tenant Learn conflict: CU security page offers CMK double encryption; Foundry CMK table marks CU "N/A" on Foundry-managed storage and requires bring-your-own storage

Answering common requests

  • "Nothing may stay in Microsoft storage." No documented CU mode removes the up-to-24-hour retention for asynchronous analysis. Options: delete results right after retrieval (grant a role that includes delete), use the synchronous preview operations for read/layout, keep Studio projects in your own storage, and choose model deployment types that match residency rules.
  • "Bring-your-own storage covers everything." It covers CU Studio project data. The account-level userOwnedStorage property is documented for Speech and Language, not for DI or CU analysis.
  • "Is the service available in our region?" Check CU region support or DI regional availability before designing around it; do not hardcode lists.
  • Logs: CU logs the analyzer name for reporting and debugging. Your own Azure Monitor diagnostic settings decide what request metadata you keep.

Verify instead of assuming

  • Role for deletion: follow access and roles and run its access checker with --operation cu-delete-result.
  • After a delete, check whether a GET on the same result ID still returns the result; record the UTC time and HTTP status, not the content.
  • Region, CMK and network settings: az cognitiveservices account show -n <name> -g <rg> --query "{location:location, encryption:properties.encryption.keySource, publicNetworkAccess:properties.publicNetworkAccess}".

Sources (retrieved 2026-11-18): DI data, privacy, security · CU data, privacy, security · CU Delete Result · CU synchronous operations · DI CMK · Foundry CMK · Foundry: connect your own storage

Source: SKILL.md on GitHub

skilld matched fixed text patterns in SKILL.md and file names. Patterns miss obfuscated code.

skilld run checks every file with the same patterns. It asks for approval before it loads a Skill with a behavior marked Needs approval.

1 warning11d4 checks · Risk SAFE
  • Gen Agent Trust Hub11d

    This skill provides tools and guidance for developing and troubleshooting Azure AI services. It includes a local diagnostic script for analyzing Speech SDK logs with robust safety controls, along with comprehensive references for official Azure SDKs and secure authentication practices.

  • Socket11d

    No alerts

  • Snyk11d

    Risk: LOW · No issues

  • Runlayer9mo

    15/15 files flagged

Signed by skilld at 0d60658. This ties the file your Agent reads to that commit on GitHub. It does not review the instructions.

Last checked against GitHub yesterday.

Activeupdated 2 days ago
metadata
{
  "author": "Microsoft",
  "version": "1.2.6"
}

README badge

README badge for microsoft/skills/azure-ai