Microsoft 365 Graph API MCP server for personal and organization accounts, with OAuth, read-only mode, scope allowlists, presets, multi-account support, and Streamable HTTP deployment.
Without `--read-only`, enabled Microsoft Graph tools can create, update, send, delete, or otherwise change Microsoft 365 data., Organization mode exposes broader work-account surfaces including Teams, SharePoint, shared mailboxes, online meetings, users, presence, and search., Use `--preset`, `--enabled-tools`, `--allowed-scopes`, and `--read-only` to keep the tool surface smaller than the default., HTTP mode requires OAuth bearer tokens for MCP requests and should be served only from trusted hosts with reviewed redirect URI allowlists., Local stdio mode can cache Microsoft tokens in the OS credential store or fallback files; protect cache paths and account pinning settings., Bring-your-own-token mode and Azure Key Vault integration require careful secret handling and rotation policies.
Privacy notes
Tool calls can expose email, calendar events, contacts, OneDrive files, Excel workbook data, OneNote pages, tasks, Planner data, Teams chats, SharePoint lists, user directory data, meeting details, and search results., OAuth tokens, client IDs, tenant IDs, client secrets, selected account metadata, token cache paths, and Key Vault URLs are sensitive configuration., Multi-account mode can reveal account usernames, account selection state, and which account handled a request., Audit logs intentionally avoid tool parameters and Graph response bodies, but still record request IDs, user principal names, tool names, HTTP method, status, duration, and error classifications., Microsoft Graph data may be governed by tenant retention, legal hold, compliance, GDPR, HIPAA, PIPEDA, SOC 2, or customer confidentiality rules.
Author
Softeria
Submitted by
oktofeesh1
Claim status
unclaimed
Last verified
2026-06-06
Decision playbook
Review trust signals before you adopt
Signals are present but mixed. Use the checklist below to confirm the source and operational safety for your environment.
Compare context
Selected
0
Current score
63
Baseline
—
Delta
No baseline selected
No major trust-signal divergence detected in the current selection.
Source and provenance checks
Needs review
Confirm ownership and provenance before trusting install instructions.
Source link availableRequired
Open the canonical repository and verify ownership.
Done
Source provenance statusRequired
Marked as source-backed.
Done
Metadata reviewed
No reviewed flag detected in metadata.
Pending
Safety and privacy checks
Complete
Validate risk disclosures before installation or API wiring.
Safety notes presentRequired
Review the listed safety guidance before running commands.
Done
Privacy notes presentRequired
Review data handling notes before connecting accounts or secrets.
Done
Trust level risk gateRequired
Trust level does not block evaluation.
Done
Package and install checks
Needs review
Check package metadata and artifact integrity signals.
Install payload available
Install or copy payload is available for review.
Done
Package verification flag
No package verification flag provided.
Pending
Checksum metadata
No checksum provided for downloaded artifact.
Pending
Compare-driven decision checks
Needs review
Use compare context to validate trade-offs before adoption.
Compare tray has multiple entries
Add at least one more entry to compare trust differences.
6 safety and 5 privacy notes across 6 risk areas. Review closely: credentials & tokens, permissions & scopes, network access.
6 areas
SafetyGeneralWithout `--read-only`, enabled Microsoft Graph tools can create, update, send, delete, or otherwise change Microsoft 365 data.
SafetyGeneralOrganization mode exposes broader work-account surfaces including Teams, SharePoint, shared mailboxes, online meetings, users, presence, and search.
SafetyPermissions & scopesUse `--preset`, `--enabled-tools`, `--allowed-scopes`, and `--read-only` to keep the tool surface smaller than the default.
SafetyCredentials & tokensHTTP mode requires OAuth bearer tokens for MCP requests and should be served only from trusted hosts with reviewed redirect URI allowlists.
SafetyCredentials & tokensLocal stdio mode can cache Microsoft tokens in the OS credential store or fallback files; protect cache paths and account pinning settings.
SafetyCredentials & tokensBring-your-own-token mode and Azure Key Vault integration require careful secret handling and rotation policies.
PrivacyLocal filesTool calls can expose email, calendar events, contacts, OneDrive files, Excel workbook data, OneNote pages, tasks, Planner data, Teams chats, SharePoint lists, user directory data, meeting details, and search results.
PrivacyNetwork accessMulti-account mode can reveal account usernames, account selection state, and which account handled a request.
PrivacyNetwork accessAudit logs intentionally avoid tool parameters and Graph response bodies, but still record request IDs, user principal names, tool names, HTTP method, status, duration, and error classifications.
PrivacyData retentionMicrosoft Graph data may be governed by tenant retention, legal hold, compliance, GDPR, HIPAA, PIPEDA, SOC 2, or customer confidentiality rules.
Disclosure: MIT-licensed community Microsoft 365 MCP server from Softeria. This entry is separate from the PnP CLI for Microsoft 365 MCP Server entry, which wraps CLI for Microsoft 365 commands instead of exposing generated Microsoft Graph tools.
Safety notes
Without `--read-only`, enabled Microsoft Graph tools can create, update, send, delete, or otherwise change Microsoft 365 data.
Organization mode exposes broader work-account surfaces including Teams, SharePoint, shared mailboxes, online meetings, users, presence, and search.
Use `--preset`, `--enabled-tools`, `--allowed-scopes`, and `--read-only` to keep the tool surface smaller than the default.
HTTP mode requires OAuth bearer tokens for MCP requests and should be served only from trusted hosts with reviewed redirect URI allowlists.
Local stdio mode can cache Microsoft tokens in the OS credential store or fallback files; protect cache paths and account pinning settings.
Bring-your-own-token mode and Azure Key Vault integration require careful secret handling and rotation policies.
Privacy notes
Tool calls can expose email, calendar events, contacts, OneDrive files, Excel workbook data, OneNote pages, tasks, Planner data, Teams chats, SharePoint lists, user directory data, meeting details, and search results.
Multi-account mode can reveal account usernames, account selection state, and which account handled a request.
Audit logs intentionally avoid tool parameters and Graph response bodies, but still record request IDs, user principal names, tool names, HTTP method, status, duration, and error classifications.
Microsoft Graph data may be governed by tenant retention, legal hold, compliance, GDPR, HIPAA, PIPEDA, SOC 2, or customer confidentiality rules.
Prerequisites
Node.js 20 or newer recommended for the current package.
Personal Microsoft account or Microsoft Entra work/school account with approved Graph access.
Microsoft Graph permissions reviewed with `--list-permissions` for the selected preset, tools, org mode, and allowed scopes.
Admin consent or tenant approval for organization tools such as Teams, SharePoint, shared mailboxes, users, presence, and meetings.
A token storage plan for local stdio mode, or OAuth configuration for Streamable HTTP deployments.
Schema details
Install type
cli
Troubleshooting
No
Source repository stats
Scope
Source repo
Collection metadata
Estimated setup
25 minutes
Difficulty
advanced
Tool listing metadata
Disclosure
MIT-licensed community Microsoft 365 MCP server from Softeria. This entry is separate from the PnP CLI for Microsoft 365 MCP Server entry, which wraps CLI for Microsoft 365 commands instead of exposing generated Microsoft Graph tools.
Softeria Microsoft 365 Graph MCP Server connects MCP clients to Microsoft 365
through Microsoft Graph. It supports personal accounts by default and
organization accounts with --org-mode, covering Outlook mail and calendar,
OneDrive files, Excel, OneNote, contacts, tasks, Planner, search, Teams,
SharePoint, users, shared mailboxes, meetings, and related Graph surfaces.
Use it when the agent needs structured Microsoft Graph tools instead of a
general Microsoft 365 command runner. Start with read-only mode and a small
preset, then expand scopes only after the intended workflow and tenant
permissions are clear.
These sources were reviewed on 2026-06-06. Prefer the live repository,
README, npm package metadata, deployment guide, security policy, package
metadata, tool category metadata, and server implementation for current setup
details.
Features
Authenticate with Microsoft through device code flow, OAuth authorization code
flow in HTTP mode, or bring-your-own-token mode.
Use personal-account tools for Outlook, calendar, OneDrive, Excel, OneNote,
To Do, Planner, contacts, profile, and search.
Enable organization mode for Teams, SharePoint, shared mailboxes, shared
calendars, online meetings, transcripts, attendance reports, users, presence,
virtual events, and work search.
Restrict the tool surface with --read-only, --preset, --enabled-tools,
--allowed-scopes, and dynamic discovery.
List required Microsoft Graph permissions before deployment.
Support Global Microsoft 365 and China 21Vianet cloud endpoints.
Serve stdio locally or Streamable HTTP with OAuth 2.1 for hosted clients.
Support multi-account use, account pinning, token cache controls, audit logs,
Graph retry/circuit-breaker behavior, and optional TOON output.
Installation
Start with a read-only, narrow preset:
npx -y @softeria/ms-365-mcp-server --read-only --preset outlook
For work or school accounts, add --org-mode after tenant permissions are
approved:
npx -y @softeria/ms-365-mcp-server --org-mode --read-only --preset teams
Before enabling broader tools, list the effective Graph permissions:
npx -y @softeria/ms-365-mcp-server --org-mode --preset teams --list-permissions
Use Cases
Summarize recent Outlook messages or calendar events from a personal account.
Retrieve OneDrive files or Excel workbook data for a task-specific context.
Search Microsoft 365 content across mail, files, notes, tasks, and work data.
Let an agent inspect Teams or SharePoint data after admin consent and org-mode
approval.
Run a hosted Streamable HTTP MCP endpoint where each user authenticates with
their own Microsoft account.
Use read-only presets to keep initial context and permissions smaller.
Safety and Privacy
Microsoft 365 data is often sensitive by default. Treat this server like access
to a real Microsoft account, not like a documentation helper. Prefer read-only
mode, exact presets, narrow enabled-tool regexes, and allowed-scope boundaries.
Review --list-permissions output before asking an admin to consent.
For production HTTP deployments, configure a dedicated Entra app, exact redirect
URI allowlists, a public URL, managed identity or Key Vault for secrets, and
audit logging. For stdio deployments, protect local token caches and pin the
expected account so the wrong mailbox or tenant is not exposed by accident.
Duplicate Notes
The catalog already includes CLI for Microsoft 365 MCP Server from the PnP
organization. That entry lets agents run CLI for Microsoft 365 commands through
an existing m365 login context. This entry covers the separate Softeria Graph
API server, npm package, OAuth modes, read-only flag, tool presets, scope
allowlists, and generated Microsoft Graph tool surface.
Show that Softeria Microsoft 365 Graph MCP Server is listed on HeyClaude. Paste this Markdown into your README — it renders the badge and links back to this page.
[](https://heyclau.de/entry/mcp/softeria-ms-365-graph-mcp-server)
How it compares
Softeria Microsoft 365 Graph MCP Server side by side with 3 alternatives on trust, install, platform support, and disclosed safety notes — all from reviewed registry metadata.
Microsoft 365 Graph API MCP server for personal and organization accounts, with OAuth, read-only mode, scope allowlists, presets, multi-account support, and Streamable HTTP deployment.
MCP server for CLI for Microsoft 365 that lets Claude use natural language to search command docs and run Microsoft 365 management commands for Entra ID, SharePoint, Teams, Outlook, OneDrive, Planner, Power Platform, and more.
Official Agentset MCP server that lets Claude retrieve cited knowledge-base results from an Agentset namespace through the `knowledge-base-retrieve` tool, with optional tenant scoping and custom tool descriptions.
✓Without `--read-only`, enabled Microsoft Graph tools can create, update, send, delete, or otherwise change Microsoft 365 data.
Organization mode exposes broader work-account surfaces including Teams, SharePoint, shared mailboxes, online meetings, users, presence, and search.
Use `--preset`, `--enabled-tools`, `--allowed-scopes`, and `--read-only` to keep the tool surface smaller than the default.
HTTP mode requires OAuth bearer tokens for MCP requests and should be served only from trusted hosts with reviewed redirect URI allowlists.
Local stdio mode can cache Microsoft tokens in the OS credential store or fallback files; protect cache paths and account pinning settings.
Bring-your-own-token mode and Azure Key Vault integration require careful secret handling and rotation policies.
✓The MCP server does not authenticate by itself; it reuses the existing CLI for Microsoft 365 authentication context.
The m365RunCommand tool can run CLI for Microsoft 365 commands that manage Entra ID, OneDrive, OneNote, Outlook, Planner, Power Apps, Power Automate, Power Platform, SharePoint Embedded, SharePoint Online, Teams, Viva Engage, and other tenant surfaces.
Natural-language prompts may result in a chain of CLI commands, so review the resolved command docs and exact proposed operations before execution.
Commands can create teams, lists, columns, items, Planner plans, Power Platform changes, flow state changes, posts, permissions, and other tenant resources.
Prefer MFA-capable interactive or browser authentication for user sessions and certificate or secret authentication only for reviewed automation scenarios.
Avoid username and password authentication when possible because the upstream best-practices guide notes MFA limitations.
✓The MCP server sends Claude's retrieval queries to the Agentset API using the configured API key and namespace.
The `knowledge-base-retrieve` tool can return up to 100 results per call and can rerank results by relevance.
Namespace and tenant selection control which indexed documents are searchable; review them before connecting a shared agent.
API keys should be scoped, rotated, and stored only in the MCP server environment or a secret manager.
Custom tool descriptions can influence when the model calls the retrieval tool, so review them before use in production workflows.
✓AI Game Developer can create, move, copy, modify, and delete Unity assets, scenes, GameObjects, components, scripts, packages, and generated project files.
Tools include dynamic C# script execution, C# reflection method lookup and calls, package installation/removal, Unity test execution, editor state changes, play mode control, screenshots, and profiler access.
The MCP server supports streamable HTTP and stdio; HTTP deployments should require a bearer token and stay bound to trusted interfaces.
Server variables include optional authorization and webhook settings; webhook endpoints can receive tool, prompt, resource, connection, and authorization events.
Unity runtime connections can expose compiled game state or in-game behavior to an MCP client, not just editor-only project data.
Use source control, backups, tool filtering, and explicit review before allowing destructive tools such as asset deletion, package removal, script execution, or reflection calls.
Privacy notes
✓Tool calls can expose email, calendar events, contacts, OneDrive files, Excel workbook data, OneNote pages, tasks, Planner data, Teams chats, SharePoint lists, user directory data, meeting details, and search results.
OAuth tokens, client IDs, tenant IDs, client secrets, selected account metadata, token cache paths, and Key Vault URLs are sensitive configuration.
Multi-account mode can reveal account usernames, account selection state, and which account handled a request.
Audit logs intentionally avoid tool parameters and Graph response bodies, but still record request IDs, user principal names, tool names, HTTP method, status, duration, and error classifications.
Microsoft Graph data may be governed by tenant retention, legal hold, compliance, GDPR, HIPAA, PIPEDA, SOC 2, or customer confidentiality rules.
✓CLI output can expose tenant IDs, user profiles, group membership, emails, calendar or Outlook data, Teams messages, SharePoint URLs, list contents, Planner tasks, flow names, app metadata, and Power Platform environment details.
Microsoft 365 tokens, certificates, client secrets, tenant IDs, site URLs, user IDs, and generated command output should stay out of prompts, issues, logs, screenshots, and committed files.
Tenant-management transcripts may contain regulated business data, employee data, customer data, or internal collaboration records.
Configure output and error handling carefully so sensitive command output is not written into persistent chat logs or debug logs unnecessarily.
✓Retrieved chunks can include private documents, product specs, policies, support content, internal procedures, historical project information, or customer-specific data.
Retrieval queries, namespace IDs, tenant IDs, document chunks, citations, and tool outputs may be visible to the MCP client, model provider, Agentset logs, and application telemetry.
Tenant IDs are useful for data segregation, but incorrect tenant or namespace configuration can expose the wrong knowledge base.
Do not paste API keys, namespace IDs, tenant IDs, or retrieved private chunks into shared issue reports, screenshots, or repository files.
✓Tool calls may expose Unity project paths, asset names, scene hierarchy, serialized object data, scripts, logs, screenshots, profiler metrics, test output, package metadata, and runtime state.
MCP config files and Unity plugin config can contain server URLs, connection modes, bearer tokens, authorization settings, enabled tool IDs, and cloud or local endpoint details.
Optional webhooks can receive tool, prompt, resource, connection, authorization, and token-bearing request data.
Screenshots from Game View, Scene View, cameras, or isolated GameObjects may include proprietary artwork, level design, UI, debug overlays, or unreleased game content.
Generated skills and AI-client configuration files can reveal available tools, project structure, installed packages, and local workflow assumptions.
Prerequisites
Node.js 20 or newer recommended for the current package.
Personal Microsoft account or Microsoft Entra work/school account with approved Graph access.
Microsoft Graph permissions reviewed with `--list-permissions` for the selected preset, tools, org mode, and allowed scopes.
Admin consent or tenant approval for organization tools such as Teams, SharePoint, shared mailboxes, users, presence, and meetings.
Node.js 20 or newer.
CLI for Microsoft 365 installed globally.
Authenticated CLI for Microsoft 365 session created with an approved m365 login method.
CLI configuration reviewed for prompt handling, output format, and detailed error help.
Agentset account or self-hosted Agentset deployment with a populated namespace.
Agentset API key with access to the namespace Claude should query.
Node.js 18.17 or newer for running the `@agentset/mcp` package.
Namespace ID selected with `--ns` or `AGENTSET_NAMESPACE_ID`.
Unity project using Unity 2022.3 or newer for the current package metadata.
Unity Hub or a local Unity Editor installation.
Node.js 20.19 or newer, or Node.js 22.12 or newer, for `unity-mcp-cli`.
An MCP client such as Claude Code, Claude Desktop, Codex, Cursor, Gemini CLI, GitHub Copilot, Cline, or another supported client.
Install
npx -y @softeria/ms-365-mcp-server --read-only --preset outlook
npx -y @pnp/cli-microsoft365-mcp-server@latest
Run `npx @agentset/mcp --ns <namespace-id>` with `AGENTSET_API_KEY` set in the MCP server environment.