# ppiova/mcp-servers-microsoft-ecosystem

Source: https://glimind.com/server/github%2Fppiova%2Fmcp-servers-microsoft-ecosystem
Registry id: `github/ppiova/mcp-servers-microsoft-ecosystem`
Status: unknown

## Tools (33)

- `append_message`: Append a message to an existing conversation. Role is typically 'user'.
- `chat`: Run a single-turn chat completion on a GitHub Model. Args: model: Model id, e.g. `openai/gpt-4.1-mini`, `microsoft/Phi-4`, `meta/Llama-3.3-70B-Instruct`. Use `l
- `compare_models`: Run the same prompt across multiple models and return all responses. Perfect for *"which model handles this best?"* workflows. Errors on a single model don't fa
- `create_conversation`: Create a new conversation in Foundry, optionally seeded with a user message. Returns `{conversation_id}`. Pair with `invoke_agent` to run a Foundry agent on the
- `get_agent`: Return metadata for a specific agent (latest version) — instructions, model, tools.
- `get_article`: Fetch a Microsoft Learn article and return its content as Markdown. Args: url: Full `https://learn.microsoft.com/...` URL. Returned by `search`. max_chars: Trun
- `get_deployment`: Get full metadata for a specific deployment — model, SKU, capacity, RAI policy, rate limits.
- `get_model_info`: Return the full catalog entry for a model (e.g. `openai/gpt-4.1`).
- `get_tenant_overview`: Snapshot counts for the tenant: users, groups, applications, service principals. Useful as a first call to understand the size and shape of a tenant before dril
- `get_user`: Return a user's profile, manager and assigned licenses. Args: upn_or_id: User principal name (`alice@contoso.com`) or object ID.
- `invoke_agent`: Run a Foundry agent on an existing conversation and return its response. Uses the OpenAI Responses API with `agent_reference` pointing at the named agent.
- `list_agents`: List every agent deployed in the configured Microsoft Foundry project.
- `list_all_deployments`: Flat view of every deployment across every OpenAI account in scope. Optionally filter by `location` (Azure region, e.g. 'eastus2'). Handy for questions like *"w
- `list_applications`: List app registrations in the tenant. Args: filter_query: OData $filter, e.g. "signInAudience eq 'AzureADMyOrg'". top: Page size (max 100).
- `list_conversation`: List items in a conversation — messages, tool calls, agent outputs.
- `list_deployments`: List model deployments on a specific Azure OpenAI account.
- `list_directory_role_assignments`: List active directory role assignments (admin roles). Returns role display names with the principals they're assigned to. Requires `RoleManagement.Read.Director
- `list_distinct_types`: Summarize the distinct ARM resource types present in the estate, with counts. Great first call when exploring an unfamiliar subscription.
- `list_group_members`: List direct members of a group (users, groups, service principals).
- `list_groups`: List groups (security + Microsoft 365), capped at 100 per call. Args: filter_query: OData $filter, e.g. "groupTypes/any(c:c eq 'Unified')" for M365 groups. top:
- `list_known_scopes`: Return a curated list of popular `scope` filter values for `search`. Handy when you want to restrict results to a specific Microsoft product area (e.g. only Azu
- `list_models`: List all models available on GitHub Models. Args: publisher: Optional filter — e.g. "OpenAI", "Microsoft", "Meta", "Mistral AI", "Cohere", "AI21 Labs", "DeepSee
- `list_openai_accounts`: List Azure OpenAI accounts (`kind == 'OpenAI'`) across subscriptions. Omit `subscription_id` to sweep all accessible subscriptions.
- `list_resource_groups`: List resource groups inside a subscription, with location and tags.
- `list_resources_by_type`: List all resources of a given ARM type (e.g. 'Microsoft.Storage/storageAccounts'). Use `list_distinct_types` first if unsure what types exist in the estate.
- `list_service_principals`: List service principals in the tenant (apps + managed identities). Args: filter_query: OData $filter, e.g. "servicePrincipalType eq 'ManagedIdentity'". top: Pag
- `list_subscriptions`: List every Azure subscription the signed-in identity can see.
- `list_usages`: Show Azure OpenAI **quota usage** for a region (tokens/min per SKU & model). Great for FinOps and capacity planning — spot which deployments are near their quot
- `list_users`: List directory users (capped at 100 per call, no pagination beyond that). Args: filter_query: OData $filter, e.g. "accountEnabled eq false" or "userType eq 'Gue
- `list_vms`: List virtual machines, optionally filtered by subscription and/or Azure region. Returns name, resource group, location, VM size, OS type, and power state where 
- `query_resources`
- `quick_ask`: One-shot: create a conversation seeded with `prompt`, invoke `agent_name`, return just the final output text.
- `search`: Search Microsoft Learn documentation. Args: query: What to search for. Plain natural-language works. locale: Language/region code — e.g. "en-us", "es-es", "pt-b

Measured by Glimind (https://glimind.com): MCP handshake + tool list only, no tool is ever called. CC BY 4.0.
