TeamsCAUTION
MCP server providing comprehensive Microsoft Teams and Graph API access for AI assistants including messaging, search, and user management.
Overview
From the repository's own README, as read at the audited commit. Badges and raw HTML are left out.
[](https://www.npmjs.com/package/@floriscornel/teams-mcp) [](https://www.npmjs.com/package/@floriscornel/teams-mcp) [](https://app.codecov.io/gh/floriscornel/teams-mcp) [](https://opensource.org/licenses/MIT) [](https://github.com/floriscornel/teams-mcp/stargazers)
A Model Context Protocol (MCP) server that provides seamless integration with Microsoft Graph APIs, enabling AI assistants to interact with Microsoft Teams, users, chats, files, and organizational data.
📦 Installation
To use this MCP server in Cursor/Claude/VS Code, add the following configuration:
{
"mcpServers": {
"teams-mcp": {
"command": "npx",
"args": ["-y", "@floriscornel/teams-mcp@latest"]
}
}
}🚀 Features
🔐 Authentication
- OAuth 2.0 device code authentication flow with Microsoft Graph
- Secure token management, cache persistence, and refresh token renewal
- Authentication status checking and logout support
- Read-only mode with reduced scopes
- Direct
AUTH_TOKENsupport for pre-issued Microsoft Graph access tokens
👥 User Management
- Get current user information
- Search users by name or email
- Retrieve detailed user profiles
- Access organizational directory data
🏢 Microsoft Teams Integration
- Teams Management
- List user's joined teams
- Access team details and metadata
- Channel Operations
- List channels within teams
- Retrieve channel messages and r
722b013e0a78OBSERVED · 2026-10-07Connect
Built from this server's own package name, version and transport as found in its source — not copied from anyone's documentation, so it cannot drift against a page we do not control. Replace the environment placeholders with a token scoped to the least it needs.
claude mcp add teams-mcp --env AUTH_TOKEN=${AUTH_TOKEN} -- npx -y @floriscornel/[email protected]{
"mcpServers": {
"teams-mcp": {
"command": "npx",
"args": [
"-y",
"@floriscornel/[email protected]"
],
"env": {
"AUTH_TOKEN": "${AUTH_TOKEN}"
}
}
}
}Exposed tools (31)
19 read · 10 write · 2 destructive. Blast radius: 2 tools can delete or overwrite — an agent that can be talked into calling a tool can be talked into calling this one.
| Tool | Risk | Description |
|---|---|---|
add_chat_member | write | |
auth_status | read | |
create_chat | write | |
delete_channel_message | destructive | |
delete_chat_message | destructive | |
download_chat_hosted_content | read | |
download_message_hosted_content | read | |
get_channel_message_replies | read | |
get_channel_messages | read | |
get_chat_messages | read | |
get_current_user | read | |
get_my_mentions | read | |
get_user | read | |
list_channels | read | |
list_chats | read | |
list_team_members | read | |
list_teams | read | |
reply_to_channel_message | read | |
search_messages | read | |
search_users | read | |
search_users_for_mentions | read | |
send_channel_message | write | |
send_chat_message | write | |
send_file_to_channel | write | |
send_file_to_chat | write | |
set_channel_message_reaction | write | |
set_chat_message_reaction | write | |
unset_channel_message_reaction | read | |
unset_chat_message_reaction | read | |
update_channel_message | write | |
update_chat_message | write |
Trust audit
CAUTIONgrade B · trust 89/100 Install with care. The audit found things worth knowing before you trust its output.
| Layer | What it checks | Result |
|---|---|---|
| L0 | Provenance & inventory | PASS |
| L1 | Static analysis of the code | WARN |
| L2 | Instruction surface (what it tells the agent) | PASS |
| L3 | Class-specific surface | WARN |
| L4 | Behavioural (sandbox) | SKIPPED |
What the source does
- Filesystem
- declared (1 observation(s))
- Network
- declared (2 observation(s))
- Shell
- none-observed
- Dependencies
- not all pinned
- Secrets in source
- none-found
Findings (11)
console.log(`⏰ Access token expires: ${expiresAt.toLocaleString()}`);delete_channel_message, delete_chat_message
import { mockUser, server } from "../../test-utils/setup.js";vi.mock("../../msal-cache.js", () => ({} from "../../test-utils/setup.js";
import type { GraphService } from "../../services/graph.js";import type { FileUploadResult } from "../../utils/file-upload.js";const payload = JSON.parse(atob(tokenSplits[1]));
@azure/identity, @azure/identity-cache-persistence, @microsoft/microsoft-graph-client, @microsoft/microsoft-graph-types, @modelcontextprotocol/sdk, @types/dompurify, @types/marked, dompurify
**Full access:**
# Full access (default)
Gates applied: no_behavioural_pass.
722b013e0a78full audit observations/trust-audit/mcp-server/floriscornel__teams.json · Report an issue / request a re-scanAudit history
Every audit this server has had. A grade with a past is a grade somebody is still checking.
| Date | Source | Verdict | Grade | Score | Change |
|---|---|---|---|---|---|
| 2026-10-07 | 722b013e0a78 | CAUTION | B | 89 | first audit |
Questions
What is the Teams MCP server?
MCP server providing comprehensive Microsoft Teams and Graph API access for AI assistants including messaging, search, and user management.
What tools does Teams expose?
31 in total: 19 read-only, 10 that write, and 2 that can delete or overwrite (delete_channel_message, delete_chat_message). Every one is listed on this page with its risk.
Is Teams safe to connect to an agent?
With care. The audit graded it B (89/100) and found 11 things worth knowing before you trust this server, listed below with the exact line each was found on. Separately from the audit: 2 of its tools can destroy data, so scope the token you give it to what you actually need.
What credentials does Teams need?
It reads AUTH_TOKEN from the environment. Give it a token scoped to the least it needs — an agent that can be talked into calling a tool can be talked into calling it with your credentials.
How does Teams run?
It speaks stdio, so it runs as a local process your client starts. It is published on npm as @floriscornel/teams-mcp at 1.0.1.
How current is this page?
The grade is for one exact copy of the source (722b013e0a78), read on 2026-10-07. The repository is watched and re-audited when it changes.