---
title: Microsoft Foundry Skill for Claude Code
canonical: https://skillselion.com/skills/microsoft/azure-skills/microsoft-foundry
content_type: text/markdown
description: "microsoft-foundry is an agent skill that runs Azure AI Foundry fine-tuning workflows for SFT, DPO, and RFT with scripts for data validation, training."
---

# Microsoft Foundry Skill

Title: Microsoft Foundry Skill for Claude Code

> microsoft-foundry is an agent skill that runs Azure AI Foundry fine-tuning workflows for SFT, DPO, and RFT with scripts for data validation, training.

Canonical: https://skillselion.com/skills/microsoft/azure-skills/microsoft-foundry

microsoft-foundry is an agent skill that runs Azure AI Foundry fine-tuning workflows for SFT, DPO, and RFT with scripts for data validation, training, deployment, and evaluation.

The microsoft-foundry skill guides fine-tuning on Azure AI Foundry with supervised fine-tuning, direct preference optimization, and reinforcement fine-tuning with graders. It spans dataset preparation, validation, training job submission, monitoring, grader calibration, deployment, evaluation, and cleanup across quickstart, full pipeline, dataset creation, iterative training, and diagnose-poor-results workflows. Bundled Python scripts cover submit_training, monitor_training, calibrate_grader, check_training curves, deploy_model via ARM REST, evaluate_model with an LLM judge, convert_dataset formats, generate distillation data, score_dataset quality, and cleanup resources. Five rules insist on baseline evaluation first, validated data before submit, RFT grader calibration targeting 25-50% base-model failure, checkpoint evaluation before deploy, and token cost measurement alongside accuracy. The skill explicitly excludes general deployment without fine-tuning, agent creation, and prompt-only optimization, routing those to sibling azure-skills modules. Error handling documents API version issues, globalStandard training types, RFT endpoint scaling, deployment race conditions, and con.

## Key facts

- Installs: 720,039
- GitHub stars: 1,647
- Author: microsoft
- Repository: microsoft/azure-skills
- Works in Grok Bot: SKILL.md frontmatter name `microsoft-foundry` and a description are valid. The instructions do not depend on Claude Code hooks, slash commands, the Task or subagent tool, or the Claude CLI. Checked 2026-11-17.
- Install command: `npx skills add https://github.com/microsoft/azure-skills --skill microsoft-foundry`
- Last updated: November 17, 2026

## What Microsoft Foundry does

Fine-tune Azure AI Foundry models with SFT, DPO, or RFT using validated datasets, training scripts, deployment, and evaluation.

How do I fine-tune an Azure AI Foundry model with the right training type, validated data, grader calibration, and safe deployment?

Submitted and monitored fine-tuning jobs with evaluated checkpoints and a deployed fine-tuned model endpoint when results pass review.

Best for: Developers training custom Azure AI Foundry models who need scripted SFT, DPO, or RFT pipelines with validation and evaluation.

Not for: Skip for general model deploy without training, agent creation, or prompt optimization without a training job.

- Supports SFT, DPO, and RFT fine-tuning with dataset prep, training, deployment, and evaluation pipelines.
- Ships Python scripts for submit, monitor, calibrate_grader, check_training, deploy, evaluate, and cleanup.
- Validators under scripts/validate cover SFT, DPO, and RFT data formats before job submission.
- Five training rules include baseline first, data validation, RFT grader calibration, and checkpoint evaluation.
- Documents platform gotchas: openai SDK version, globalStandard OSS models, RFT S2+ scaling, and deployment races.

## SKILL.md

## Microsoft Foundry Skill

This skill helps developers work with Microsoft Foundry resources, covering model discovery and deployment, complete dev lifecycle of AI agent, evaluation workflows, and troubleshooting.

### Pre-Execution Requirements

Follow each applicable subsection below before starting its corresponding action or workflow.

#### Dependency Check and Setup

**MANDATORY:** As the first step after this skill loads, run the dependency check and setup script below from this skill's root and wait for it to finish before continuing. The script checks first and installs only missing dependencies; it does not reinstall dependencies that are already available.

**You MUST complete this check before reading or entering any sub-skill, workflow, or workflow-specific reference.**

```bash
./scripts/check-and-setup-dependencies.sh     # macOS / Linux
./scripts/check-and-setup-dependencies.ps1    # Windows (pwsh)
```

Strictly follow the script output for subsequent actions.

#### Workflow Guidance

**MANDATORY:** Before executing ANY workflow-specific steps, you MUST read the corresponding sub-skill document. Do not call workflow-specific MCP tools for a workflow without reading its skill document. This applies even if you already know the MCP tool parameters  -  the skill document contains required workflow steps, pre-checks, and validation logic that must be followed. This rule applies on every new user message that triggers a different workflow, even if the skill is already loaded.

#### Foundry MCP

**MANDATORY:** Before using Foundry MCP operations, call the Azure MCP `foundry` tool and inspect the available Foundry MCP tools and related parameters. Treat this as the discovery/help step for MCP-based workflows.

#### azd

**MANDATORY:** Before executing ANY azd command, you MUST read [azd-guidance](foundry-agent/azd-guidance/azd-guidance.md) and strictly follow the shared rules defined in it, especially the `AZURE_DEV_USER_AGENT` setting rules.

### Sub-Skills

This skill includes specialized sub-skills for specific workflows. **When a sub-skill matches the task, strictly follow its workflow:**

| Sub-Skill | When to Use | Reference |
|-----------|-------------|-----------|
| **deploy** | Deploy hosted agents to Foundry, smoke-test a deployment, create or update prompt agents, and manage agent versions and multi-environment deploys. | [deploy](foundry-agent/deploy/deploy.md) |
| **cicd** | Set up a CI/CD deployment pipeline for a Foundry agent. | [cicd](foundry-agent/cicd/cicd.md) |
| **invoke** | Send messages to an agent, single or multi-turn conversations | [invoke](foundry-agent/invoke/invoke.md) |
| **routine** | Schedule or event-trigger Foundry agents with routines; use `azd` for CRUD, enable/disable, manual dispatch, and viewing past runs, or define routines in `azure.yaml`. | [routine](foundry-agent/routine/routine.md) |
| **invocations-ws** | Build, deploy, and connect to hosted agents that speak the `invocations_ws` duplex WebSocket protocol  -  voice agents, real-time streams, and signaling for out-of-band media transports. | [invocations-ws](foundry-agent/invocations-ws/invocations-ws.md) |
| **observe** | Evaluate agent quality, run batch evals, analyze failures, optimize prompts, improve agent instructions, compare versions, set up CI/CD monitoring, and enable continuous production evaluation | [observe](foundry-agent/observe/observe.md) |
| **insights** | Pull generated agent insights, evidence, and recommendations from an existing monitor; read-only retrieval, not a new analysis run | [insights](foundry-agent/insights/insights.md) |
| **trace** | Query traces, analyze latency/failures, correlate eval results to specific responses via App Insights `customEvents` | [trace](foundry-agent/trace/trace.md) |
| **troubleshoot** | View hosted agent logs, query telemetry, diagnose failures | [troubleshoot](foundry-agent/troubleshoot/troubleshoot.md) |
| **validate** | Use only when the user explicitly asks to use this validation sub-skill or to validate Microsoft Foundry hosted-agent code against best practices. Never invoke it proactively or add it to another workflow. | [validate](foundry-agent/validate/validate.md) |
| **create (quick start)** | Create a new hosted Foundry agent from scratch end-to-end  -  scaffold, provision or use an existing Foundry project, deploy, and smoke-test. Do not use for any work on existing code. For anything not covered by the quickstart, use **create**. | [create/quick-start-hosted.md](foundry-agent/create/quick-start-hosted.md) |
| **create** | Use when the standard end-to-end happy path (quick start) doesn't fit. Create a new Foundry agent, update code of an existing agent, continue development of an existing agent, wire connections at scaffold time, use advanced setup or A2A (Agent2Agent), or recover from a failed quickstart run. | [create](foundry-agent/create/create-hosted.md) |
| **create managed harness** | Create or continue developing a Prompt Agent that explicitly uses the Foundry-managed GitHub Copilot harness. Uses `azd` and `azure.yaml`, not the ordinary Prompt Agent MCP workflow. | [create-managed-harness](foundry-agent/create/create-managed-harness.md) |
| **agent-optimizer** | Make existing Python hosted-agent code optimization-ready, configure eval.yaml, run Agent Optimizer jobs, apply candidates locally, and deploy through azd after review. | [agent-optimizer](foundry-agent/agent-optimizer/agent-optimizer.md) |
| **eval-datasets** | Harvest production traces into evaluation datasets, manage dataset versions and splits, track evaluation metrics over time, detect regressions, and maintain full lineage from trace to deployment. Use for: create dataset from traces, dataset versioning, evaluation trending, regression detection, dataset comparison, eval lineage. | [eval-datasets](foundry-agent/eval-datasets/eval-datasets.md) |
| **project/create** | Creating a new Microsoft Foundry project for hosting agents and models. Use when onboarding to Foundry or setting up new infrastructure. | [project/create/create-foundry-project.md](project/create/create-foundry-project.md) |
| **resource/create** | Creating Azure AI Services multi-service resource (Foundry resource) using Azure CLI. Use when manually provisioning AI Services resources with granular control. | [resource/create/create-foundry-resource.md](resource/create/create-foundry-resource.md) |
| **private-network** | Answer questions about Foundry network isolation **and** deploy Foundry with VNet isolation (BYO VNet, Managed VNet, hybrid). Covers architecture concepts, template selection, deployment, and post-deployment validation. | [resource/private-network/private-network.md](resource/private-network/private-network.md) |
| **models/deploy-model** | Unified model deployment with intelligent routing. Handles quick preset deployments, fully customized deployments (version/SKU/capacity/RAI), and capacity discovery across regions. Routes to sub-skills: `preset` (quick deploy), `customize` (full control), `capacity` (find availability). | [models/deploy-model/SKILL.md](models/deploy-model/SKILL.md) |
| **quota** | Managing quotas and capacity for Microsoft Foundry resources. Use when checking quota usage, troubleshooting deployment failures due to insufficient quota, requesting quota increases, or planning capacity. | [quota/quota.md](quota/quota.md) |
| **rbac** | Managing RBAC permissions, role assignments, managed identities, and service principals for Microsoft Foundry resources. Use for access control, auditing permissions, and CI/CD setup. | [rbac/rbac.md](rbac/rbac.md) |
| **finetuning** | Fine-tune models on Microsoft Foundry  -  SFT distillation, DPO preference optimization, RFT with graders and tool calling. Dataset preparation, grader calibration, training, checkpoint selection, deployment, evaluation. Use for: fine-tune, SFT, DPO, RFT, training data, grader, distillation, fine-tuned model, large file upload. | [finetuning/SKILL.md](finetuning/SKILL.md) |
| **azd-guidance** | Provide shared azd knowledge and guidance for managing Foundry agents. Read this first for any workflows related to azd. | [azd-guidance](foundry-agent/azd-guidance/azd-guidance.md) |

> 💡 **Tip:** For a complete onboarding flow: `project/create` (public) or `private-network` (VNet isolation) → `models/deploy-model` → agent workflows (`create` → `deploy` → `invoke`).

> 💡 **Fine-Tuning:** Use `finetuning` for all model customization  -  SFT distillation, DPO preference optimization, and RFT with graders. Includes quickstart, grader calibration, and training curve analysis.

> 💡 **Model Deployment:** Use `models/deploy-model` for all deployment scenarios  -  it intelligently routes between quick preset deployment, customized deployment with full control, and capacity discovery across regions.

> 💡 **Prompt Optimization:** For requests like "optimize my prompt" or "improve my agent instructions," load [observe](foundry-agent/observe/observe.md) and use the `prompt_optimize` MCP tool through that eval-driven workflow.

### Infrastructure Lifecycle

Match user intent to the correct infrastructure workflow.

| User Intent | Workflow |
|-------------|---------|
| "Create Foundry" / "Set up Foundry" (ambiguous) | Use `AskUserQuestion`: (a) just an AI Services resource, (b) a project with public access, or (c) a project with network isolation? Route: (a) → [resource/create](resource/create/create-foundry-resource.md), (b) → [project/create](project/create/create-foundry-project.md), (c) → [private-network](resource/private-network/private-network.md) |
| Set up Foundry with VNet isolation | [private-network](resource/private-network/private-network.md) |
| Create a Foundry project (public) | [project/create](project/create/create-foundry-project.md) |
| Create a bare Foundry resource | [resource/create](resource/create/create-foundry-resource.md) |

### Agent Development Lifecycle

Match user intent to the correct agent workflow. Read each sub-skill in order before executing.

| User Intent | Workflow (read in order) |
|-------------|------------------------|
| Create or develop an explicitly requested Managed Harness Agent / GitHub Copilot harness agent | [dependency check and setup](#dependency-check-and-setup) → [azd-guidance](foundry-agent/azd-guidance/azd-guidance.md) → [create-managed-harness](foundry-agent/create/create-managed-harness.md) → [deploy](foundry-agent/deploy/deploy.md) → [invoke](foundry-agent/invoke/invoke.md) |
| Create a new hosted agent end-to-end (scaffold + deploy + test) | [dependency check and setup](#dependency-check-and-setup) → [azd-guidance](foundry-agent/azd-guidance/azd-guidance.md) → [quick-start-hosted](foundry-agent/create/quick-start-hosted.md) (self-contained end-to-end) |
| Anything beyond the standard quickstart (existing code, migration, re-hosting, deployment customization, scaffold-time connections, A2A (Agent2Agent), recovery) | [dependency check and setup](#dependency-check-and-setup) → [azd-guidance](foundry-agent/azd-guidance/azd-guidance.md) → [create](foundry-agent/create/create-hosted.md) → [deploy](foundry-agent/deploy/deploy.md) → [invoke](foundry-agent/invoke/invoke.md) |
| Optimize existing Python hosted agent | [dependency check and setup](#dependency-check-and-setup) → [azd-guidance](foundry-agent/azd-guidance/azd-guidance.md) → [agent-optimizer](foundry-agent/agent-optimizer/agent-optimizer.md) → scaffold/review → eval.yaml → optimize → apply candidate → deploy → invoke |
| Deploy an agent (code already exists) | [dependency check and setup](#dependency-check-and-setup) → [azd-guidance](foundry-agent/azd-guidance/azd-guidance.md) → deploy (includes eval-suite setup) → invoke → observe (evaluate/optimize) |
| Update/redeploy an agent after code changes | [dependency check and setup](#dependency-check-and-setup) → [azd-guidance](foundry-agent/azd-guidance/azd-guidance.md) → deploy (includes eval-suite setup) → invoke → observe (evaluate/optimize) |
| Set up a CI/CD deployment pipeline for a hosted agent | [dependency check and setup](#dependency-check-and-setup) → [azd-guidance](foundry-agent/azd-guidance/azd-guidance.md) → cicd |
| Invoke/test/chat with an agent | [dependency check and setup](#dependency-check-and-setup) → [azd-guidance](foundry-agent/azd-guidance/azd-guidance.md) → invoke |
| Schedule/event-trigger an agent, or CRUD/enable/disable/dispatch a routine | [dependency check and setup](#dependency-check-and-setup) → [azd-guidance](foundry-agent/azd-guidance/azd-guidance.md) → routine |
| Optimize / improve agent prompt or instructions | observe (Step 4: Optimize) |
| Evaluate and optimize agent (full loop) | observe |
| Enable continuous evaluation monitoring | observe (Step 6: CI/CD & Monitoring) |
| Pull agent insights / list generated issues and recommendations | [dependency check and setup](#dependency-check-and-setup) → [insights](foundry-agent/insights/insights.md) (all pages with expanded evidence; read-only) |
| Troubleshoot an agent issue | [dependency check and setup](#dependency-check-and-setup) → [azd-guidance](foundry-agent/azd-guidance/azd-guidance.md) → invoke → troubleshoot |
| Fix a broken agent (troubleshoot + redeploy) | [dependency check and setup](#dependency-check-and-setup) → [azd-guidance](foundry-agent/azd-guidance/azd-guidance.md) → invoke → troubleshoot → apply fixes → deploy → invoke |

### Agent: .foundry Workspace Standard

Every agent source folder can keep Foundry-specific cache and overlay state under `.foundry/`:

```text
<agent-root>/
  .foundry/
    agent-metadata.yaml
    agent-metadata.prod.yaml
    suites/
    datasets/
    evaluators/
    results/
```

- In azd projects, derive deployment context (project endpoint, agent name/version, ACR, App Insights) from `azure.yaml` plus `azd env get-values`; do not duplicate those values in metadata when azd already provides them.
- `agent-metadata.yaml` is the preferred local/dev overlay for non-azd values, remote Foundry suite references, local cache paths, result summaries, and explicit overrides. Optional sidecar files such as `agent-metadata.prod.yaml` can hold a single prod or CI-targeted overlay without mixing multiple environments in one file.
- `suites/`, `datasets/`, and `evaluators/` are local cache folders. Reuse them when they are current, and ask before refreshing or overwriting them.
- See [Agent Metadata Contract](references/agent-metadata-contract.md) for the canonical schema and workflow rules.

### Agent: Setup References

- [Standard Agent Setup](references/standard-agent-setup.md)  -  advanced setup for production workloads that need data-residency control (bring-your-own Cosmos DB / Storage / AI Search via a Foundry capability host). The default `azd ai agent` flow uses **Basic Agent Setup** and does **not** provision `capabilityHosts/agents`  -  do not flag its absence as a bug. For default post-provision state, see the "Expected env-var fingerprint" section in [foundry-agent/create/create-hosted.md](foundry-agent/create/create-hosted.md).

### Agent: Common Project Context Resolution

Agent skills should run this step **only when they need configuration values they don't already have**. If a value (for example, agent root, environment, project endpoint, or agent name) is already known from the user's message or a previous skill in the same session, skip resolution for that value.

#### Step 1: Discover Agent Roots and azd Context

First check whether the workspace has `azure.yaml` with services using `host: azure.ai.agent`.

- **One azd agent service** -> classify it from `kind` and `harness`, then use that service's `project` folder or the `azure.yaml` directory as the agent root.
- **Multiple azd agent services** -> require the user to choose the target service/folder.
- **No azd agent service** -> search the workspace for `.foundry/` folders that contain `agent-metadata.yaml` or `agent-metadata.<env>.yaml`.

Excerpt. The rest of the file is on the canonical page and in the source repository.

## Deep dive: how does microsoft-foundry actually work?

We read the skill at commit `968f7bf` of microsoft/azure-skills on November 14, 2026, counting the `skills/microsoft-foundry/` tree; the repo ships a byte-identical second copy under `.github/plugins/azure-skills/skills/`, which is not counted twice. One honest note: the skill ships 231 files and our file panel carries a subset of them, so the walkthrough below quotes the repository directly.

## microsoft-foundry at a glance

1. **A cloud platform's operating manual, shipped as a skill**: twenty-four sub-skills behind one router.
2. **It runs an installer the moment it loads**, by mandate, and tags Microsoft's telemetry while it is at it.
3. **The trigger is a routing contract**, down to an explicit DO-NOT-USE list naming other skills.
4. **Documentation outranks what the agent already knows**, even correct MCP tool parameters.
5. **Every intent has a prescribed reading order**, and ambiguity routes to a question, not a guess.
6. **It is versioned like software**, because for Microsoft it is one.

## 1. A cloud platform's operating manual, shipped as a skill

The root file says what it covers in one sentence:

> This skill helps developers work with Microsoft Foundry resources, covering model discovery and deployment, complete dev lifecycle of AI agent, evaluation workflows, and troubleshooting.
> Source: SKILL.md, [overview](https://github.com/microsoft/azure-skills/blob/968f7bfaa01672e782dae2c6bab4fb9505ae047a/skills/microsoft-foundry/SKILL.md#L12)

Behind that sentence sits a table of twenty-four sub-skills, each its own document: deploy, CI/CD, invoke, routines, WebSocket agents, observability, insights, tracing, troubleshooting, validation, four flavors of create (quick start, advanced, managed harness and voice), an agent optimizer, eval datasets, project and resource provisioning, private networking, model deployment, quota, RBAC, fine-tuning, and shared azd guidance. The root SKILL.md is a 310-line router over all of it, the same index-plus-references shape we saw in [vercel-react-best-practices](/skills/vercel-labs/agent-skills/vercel-react-best-practices#deep-dive), scaled up to an entire cloud product.

## 2. It runs an installer the moment it loads

The first pre-execution requirement is unambiguous:

> **MANDATORY:** As the first step after this skill loads, run the dependency check and setup script below from this skill's root and wait for it to finish before continuing. The script checks first and installs only missing dependencies [...]
> Source: SKILL.md, ["Dependency Check and Setup"](https://github.com/microsoft/azure-skills/blob/968f7bfaa01672e782dae2c6bab4fb9505ae047a/skills/microsoft-foundry/SKILL.md#L20)

We read the script. It refuses to install `azd` itself (it exits with a pointer to Microsoft's install page), but if `azd` exists, it loops over two Microsoft azd extensions, Foundry and evaluations, and installs whichever is missing without asking (the script is idempotent: already-installed dependencies are left alone):

> for extension_id in microsoft.foundry azure.ai.evaluations; do
> [...]
>   azd extension install "$extension_id"
> Source: scripts/check-and-setup-dependencies.sh, [install step](https://github.com/microsoft/azure-skills/blob/968f7bfaa01672e782dae2c6bab4fb9505ae047a/skills/microsoft-foundry/scripts/check-and-setup-dependencies.sh#L17-L24)

The script also sets `AZURE_DEV_USER_AGENT=microsoft_foundry_skill`, and the skill's mandatory azd guidance requires that same variable inline on every azd command, so Microsoft's telemetry can attribute the traffic to this skill. Neither behavior is hidden, both are in a 27-line script anyone can read, and both are exactly the kind of thing you want to know before an agent runs it. It is the same class of finding as [find-skills' `-y` flag](/skills/vercel-labs/skills/find-skills#deep-dive): the install path is smoother than the consent path.

## 3. The trigger is a routing contract

The description is a keyword manifest, and it ends by naming what it refuses to handle:

> DO NOT USE FOR: Azure Functions, App Service, general Azure deploy (use azure-deploy), general Azure prep (use azure-prepare).
> Source: SKILL.md, [frontmatter description](https://github.com/microsoft/azure-skills/blob/968f7bfaa01672e782dae2c6bab4fb9505ae047a/skills/microsoft-foundry/SKILL.md#L3)

A negative trigger list is rarer and more useful than the positive one: it stops the skill from swallowing every Azure task in the repo and hands the general cases to sibling skills by name. Where [agent-browser](/skills/vercel-labs/agent-browser#deep-dive) writes trigger words to win tasks, microsoft-foundry also writes them to decline tasks.

## 4. Documentation outranks what the agent already knows

The sharpest rule in the file anticipates an overconfident agent:

> Do not call workflow-specific MCP tools for a workflow without reading its skill document. This applies even if you already know the MCP tool parameters  -  the skill document contains required workflow steps, pre-checks, and validation logic that must be followed.
> Source: SKILL.md, ["Workflow Guidance"](https://github.com/microsoft/azure-skills/blob/968f7bfaa01672e782dae2c6bab4fb9505ae047a/skills/microsoft-foundry/SKILL.md#L33)

That "even if you already know the parameters" is written by someone who has watched a model skip the runbook because the API call looked obvious. On a platform where a wrong deploy step costs real money and real quota, the file makes the runbook non-optional.

## 5. Every intent has a prescribed reading order

The bottom half of the file is intent tables: user phrasing on the left, an ordered chain of sub-skill documents on the right, dependency check first, azd guidance second, then the workflow. And when the user's request is genuinely ambiguous, the file says to ask instead of guess:

> "Create Foundry" / "Set up Foundry" (ambiguous) | Use `AskUserQuestion`: (a) just an AI Services resource, (b) a project with public access, or (c) a project with network isolation?
> Source: SKILL.md, ["Infrastructure Lifecycle"](https://github.com/microsoft/azure-skills/blob/968f7bfaa01672e782dae2c6bab4fb9505ae047a/skills/microsoft-foundry/SKILL.md#L88)

Three different infrastructures hide behind the words "set up Foundry", and the table refuses to let the agent pick one silently. It is the same decisions-belong-to-the-user principle [grill-me](/skills/mattpocock/skills/grill-me#deep-dive) builds a whole interview around, applied here to a single provisioning fork.

## 6. It is versioned like software

The frontmatter carries `version: "1.2.35"` under an MIT license, and the sub-skills are modules behind a router whose third mandate is to inspect the Azure MCP foundry toolset before calling any of its tools. Microsoft is treating agent instructions as a shipped, versioned product surface, not a README: the same conclusion the [react-best-practices](/skills/vercel-labs/agent-skills/vercel-react-best-practices#deep-dive) build pipeline reaches by compiling its rules.

**Tools it drives:** `azd` (with the Foundry and evaluations extensions it installs), the Azure MCP `foundry` toolset, and shell scripts shipped in the bundle.

**Behavior change, before and after:**

| You ask | Without microsoft-foundry | With microsoft-foundry |
|---|---|---|
| "Deploy my agent to Foundry" | Improvised azd calls, no pre-checks | Dependency script → azd guidance → the deploy sub-skill's full runbook |
| "Set up Foundry" | The agent picks an infrastructure shape for you | It asks which of the three setups you actually meant |

**The fine print worth knowing:**

- Loading the skill can lead to software being installed: the mandated script adds the `microsoft.foundry` and `azure.ai.evaluations` azd extensions unprompted when they are missing, and the azd guidance stamps telemetry with the skill's user agent. Read the 27-line script once; it is short and honest.
- This is Microsoft operating Microsoft: every runbook optimizes for the Foundry happy path, and the DO-NOT-USE list hands you to other Microsoft skills, not to alternatives.

## Should you install microsoft-foundry?

Install it if you run agents or models on Microsoft Foundry; it is the most operationally complete skill we have read, and the read-the-runbook-first discipline is what you want next to a cloud bill. Skip it if you are not on Foundry, and know that loading it on a machine with `azd` but without the Foundry or evaluations extension will install the missing one as a side effect of the first task.

```
npx skills add https://github.com/microsoft/azure-skills --skill microsoft-foundry
```

---

*Previous: [how lark-doc actually works](/skills/larksuite/cli/lark-doc#deep-dive). All dives: [the series index](/deep-dives).*

## Microsoft Foundry FAQ

### Does microsoft-foundry work without the Azure Developer CLI?

No. The mandated setup script exits immediately if `azd` is missing, pointing at Microsoft's install page. With `azd` present, it installs the `microsoft.foundry` and `azure.ai.evaluations` extensions itself when they are missing, and proceeds.

### Why does one skill contain twenty-four sub-skills?

Because it covers a platform, not a task. The root SKILL.md is a router: intent tables map each request (deploy, evaluate, fine-tune, quota, RBAC, networking) to an ordered chain of sub-skill documents, so the agent loads only the runbook the task needs.

### Who is microsoft-foundry for?

Developers running Azure AI Foundry fine-tuning who need SFT, DPO, or RFT scripts, validators, and deployment guidance.

### When should I use microsoft-foundry?

When preparing training data, submitting or monitoring fine-tuning jobs, calibrating RFT graders, or deploying and evaluating fine-tuned checkpoints.

### Is microsoft-foundry safe to install?

Review the Security Audits panel on this page before installing in production.

## Machine access

- [Sitemap](https://skillselion.com/sitemap.xml)
- [llms.txt](https://skillselion.com/llms.txt)
- [llms-full.txt](https://skillselion.com/llms-full.txt)
- [Developer resources](https://skillselion.com/developers)
- [OpenAPI spec](https://skillselion.com/openapi.json)
- [MCP handshake](https://skillselion.com/.well-known/mcp)
- [API catalog](https://skillselion.com/.well-known/api-catalog)
- [AI catalog](https://skillselion.com/.well-known/ai-catalog.json)
- [Pricing](https://skillselion.com/pricing.md)
- [Auth](https://skillselion.com/auth.md)
- [AGENTS.md](https://skillselion.com/AGENTS.md)
- [Contact](https://skillselion.com/contact)
- [Home](https://skillselion.com/)
