Spec Driven Development
MCP Server that facilitates spec-driven development workflows, not just Vibe Coding.
Description from the maker (original English)
Inspect, validate, upload, and submit Microsoft Forms responses
These three are never merged into one score: opinions, measurements and popularity are different things.
We have not written a plain-language summary for this one yet. Below are the facts we read from the registry.
Pick the AI app you use, then copy the settings.
MSFORMS_PROFILE_DIRECTORY, MSFORMS_BROWSER_CHANNEL, MSFORMS_HEADLESSYou need
Which app do you use?
Pick your app above. We remember it, so next time the button says “Install in …”.
Open Claude Desktop → Settings → Developer → Edit Config, and paste this into claude_desktop_config.json
{
"mcpServers": {
"msforms-api": {
"command": "npx",
"args": [
"-y",
"msforms-api",
"msforms-api-mcp"
],
"env": {
"MSFORMS_PROFILE_DIRECTORY": "your-value",
"MSFORMS_BROWSER_CHANNEL": "your-value",
"MSFORMS_HEADLESS": "your-value"
}
}
}
}Run this in your terminal
claude mcp add msforms-api --env MSFORMS_PROFILE_DIRECTORY=your-value --env MSFORMS_BROWSER_CHANNEL=your-value --env MSFORMS_HEADLESS=your-value -- npx -y msforms-api msforms-api-mcpFill in the empty values before running it.
Paste into .cursor/mcp.json in your project, or ~/.cursor/mcp.json
{
"mcpServers": {
"msforms-api": {
"command": "npx",
"args": [
"-y",
"msforms-api",
"msforms-api-mcp"
],
"env": {
"MSFORMS_PROFILE_DIRECTORY": "your-value",
"MSFORMS_BROWSER_CHANNEL": "your-value",
"MSFORMS_HEADLESS": "your-value"
}
}
}
}Your browser asks to open the app, then the app shows the settings and asks you to confirm. Key values are left empty for you to fill in.
Paste into .vscode/mcp.json in your project
{
"servers": {
"msforms-api": {
"type": "stdio",
"command": "npx",
"args": [
"-y",
"msforms-api",
"msforms-api-mcp"
],
"env": {
"MSFORMS_PROFILE_DIRECTORY": "your-value",
"MSFORMS_BROWSER_CHANNEL": "your-value",
"MSFORMS_HEADLESS": "your-value"
}
}
}
}Your browser asks to open the app, then the app shows the settings and asks you to confirm. Key values are left empty for you to fill in.
This one runs on your computer, and ChatGPT can only connect to servers on the internet. Pick another app.
Your choice stays on this device in a cookie; no account needed.
MSFORMS_PROFILE_DIRECTORY envMSFORMS_BROWSER_CHANNEL envMSFORMS_HEADLESS envThese settings are generated from the maker’s registry entry. If the README says otherwise, follow the README.
Runs on your computer; the tool list appears in your AI app after you install it.
We check information anyone can see; we have not read the code. Checks that do not apply are left out of the count. · Checked Nov 7, 2026
The registry name was verified with the GitHub account Maxim-Mazurok, which owns the repo
We have not read this repo from GitHub yet
We have not read this repo from GitHub yet
This one runs on your computer; there is no internet endpoint
Runs on your computer and uses local settings instead
Runs on your computer; we can only read tool lists from internet endpoints
Compared with every active server in this category in the registry. Stars count only when the publisher owns the repo.
It passed 1 of 1 applicable basic checks. We have not read the code. Before adding a key or token, see what access the server asks for and use a key with only the permissions you need.
The registry does not list a required key, though some servers ask you to sign in on first use.
It runs on your computer. Your AI app starts it when needed, and data does not pass through the maker’s servers (unless it calls an online service itself).
The GitHub repo last changed –, and the latest version (0.2.2) reached the registry 3 months ago.
Apps that support MCP, such as Claude Desktop, Claude Code, Cursor and VS Code. You also need Node.js on your computer.
In the same category: Spec Driven Development, Remote Desktop Commander, Intervals MCP
Servers in the same category from other makers, sorted by stars.
Reading a card:★ GitHub starsiThree pillars: reviews · checks · usageiUpdated = latest version published to the registryi
MCP Server that facilitates spec-driven development workflows, not just Vibe Coding.
Remote Desktop Commander: use files, terminal and processes on your own computers from any AI app.
Remote MCP server for intervals.icu run data and analysis
Copied from the maker’s repo as written. PopNx MCP has not reviewed or endorsed it. HTML, images and scripts are removed. Links lead off this site.
AI-native Microsoft Forms automation for inspecting forms, resolving branching, validating answers, uploading files, submitting responses, and saving response links.
Use it as:
[!NOTE]
This project was generated by GitHub Copilot App using GPT-5.6 Sol with
Extra High reasoning in Autopilot mode.
[!WARNING]
This project uses reverse-engineered Microsoft Forms web APIs. Microsoft does
not document or guarantee these endpoints. They can change without notice.
This project is not affiliated with or endorsed by Microsoft.
The CLI skill teaches agents how to inspect branching, prepare and validate answers, upload authorized files, and require explicit approval before submission:
npx -y skills add Maxim-Mazurok/msforms-api
The installer detects supported agents such as Claude Code, Cursor, GitHub Copilot, and Codex, then asks where to install the skill. Choose global scope to use it across projects.
The skill runs msforms-api through npx, so it does not require a global package installation or MCP configuration. It is an alternative interface to the MCP server; both use the same CLI and SDK behavior.
Click an install badge above, or use this MCP configuration. For Cursor, save it as ~/.cursor/mcp.json:
{
"mcpServers": { "microsoft-forms": { "command": "npx", "args": ["-y", "-p", "msforms-api@latest", "msforms-api-mcp"] } } }
On first use, ask your agent to authenticate Microsoft Forms. A browser opens for sign-in, then future operations reuse the persistent session.
Try:
Inspect this Microsoft Form and explain every question, validation, upload
rule, and branch:
https://forms.cloud.microsoft/...
The agent can prepare and validate answers without submitting. Submission uses a separate destructive MCP tool and requires explicit confirmation.
VS Code CLI installation
macOS / Linux:
code --add-mcp '{"name":"microsoft-forms","command":"npx","args":["-y","-p","msforms-api@latest","msforms-api-mcp"]}'
Windows PowerShell:
code --add-mcp '{"name":"microsoft-forms","command":"npx","args":["-y","-p","msforms-api@latest","msforms-api-mcp"]}'
Claude Desktop
Add the MCP configuration above to claude_desktop_config.json.
Claude Code
claude mcp add microsoft-forms -- npx -y -p msforms-api@latest msforms-api-mcp
Windsurf and other MCP clients
Add the MCP configuration above to your client's MCP configuration file.
| Tool | Behavior |
|---|---|
forms_authenticate | Opens visible browser login |
forms_inspect | Returns normalized and raw form structure |
forms_validate | Resolves branches and validates without submitting |
forms_upload_file | Uploads a local file and returns answer metadata |
forms_submit | Creates a response after elicited confirmation |
forms_save_response | Saves an existing response link |
forms_list_saved_responses | Lists saved response links |
forms_submit is marked destructive and non-idempotent. It uses MCP v2 multi-round-trip elicitation. Declined or false confirmation never reaches submission code.
Cross-cutting workflow and safety guidance is also available through:
npx -y -p msforms-api@latest msforms-api guide
Run without installing globally:
# Sign in once
npx -y -p msforms-api@latest msforms-api auth
npx -y -p msforms-api@latest msforms-api inspect \ --form 'https://forms.cloud.microsoft/Pages/ResponsePage.aspx?id=...'
npx -y -p msforms-api@latest msforms-api validate \ --form 'https://forms.cloud.microsoft/Pages/ResponsePage.aspx?id=...' \ --answers @answers.json
npx -y -p msforms-api@latest msforms-api submit \ --form 'https://forms.cloud.microsoft/Pages/ResponsePage.aspx?id=...' \ --answers @answers.json \ --save-response \ --confirm
Frequent CLI users can install globally:
npm install --global msforms-api
msforms-api auth
Every command emits JSON. Add --compact for single-line output.
Authentication opens a visible browser and stores its profile under:
~/.msforms-api/browser-profile
The SDK uses that profile to obtain scoped Forms cookies, an anti-forgery token, and session metadata. It does not print or return those credentials. Normal reads and submissions use direct Forms HTTP requests. File upload is browser-assisted because the Forms frontend manages SharePoint authorization.
Override the profile or browser:
msforms-api --profile ~/.config/my-forms-profile --browser chrome auth
MCP environment variables:
| Variable | Purpose |
|---|---|
MSFORMS_PROFILE_DIRECTORY | Persistent browser profile override |
MSFORMS_BROWSER_CHANNEL | chrome or msedge |
MSFORMS_HEADLESS | Set to false to show operational browser windows |
Delete ~/.msforms-api/browser-profile to remove the local session.
Answers are keyed by exact question ID returned from forms_inspect.
| Question type | Input |
|---|---|
| Text | String |
| Single choice | Option label or ID |
| Multiple choice | Array of option labels or IDs |
| Date | YYYY-MM-DD string |
| Rating or NPS | Number |
| Ranking | Ordered array containing every option label or ID |
| Matrix row | Option label, or { "id": ..., "key": ... } |
| File upload | Array of objects returned by forms_upload_file |
Matrix answers use each row question ID, not the matrix group ID.
{
"text-question-id": "Ada Lovelace", "multiple-choice-question-id": ["Desktop", "Laptop"], "date-question-id": "2026-08-12", "ranking-question-id": ["Branching", "Uploads", "Ratings"], "matrix-row-question-id": { "label": "Agree" }, "file-question-id": [ { "name": "document.pdf", "uploadSessionUrl": "https://...", "link": "https://...", "badgerToken": null, "time": 1786492949627, "status": 3, "id": "sharepoint-item-id" } ] }
npm install msforms-api
import { MicrosoftFormsClient } from "msforms-api";
const client = new MicrosoftFormsClient(); const form = await client.getForm(formUrl);
const answers = { [form.questions[0].id]: "Example answer", };
const validation = await client.validateAnswers(formUrl, answers); if (!validation.valid) { throw new Error(JSON.stringify(validation.issues)); }
const submission = await client.submitResponse(formUrl, answers, { saveResponse: true, });
if (submission.saveResponseStatus === "failed") { console.error(submission.saveResponseError); await client.saveResponse( formUrl, submission.responseId, submission.submitDate, ); }
submitResponse() is direct at SDK level. Applications using the SDK must implement their own user confirmation boundary. CLI requires --confirm; MCP requires elicitation.
submissionStatus: "submitted" or a returned responseId means the permanent response exists. A requested response-link save can fail independently. In that case, saveResponseStatus is "failed", submission details remain in the result, and callers must not submit again. Retry only the idempotent saveResponse() operation.
The normalized form model includes:
branchTargetId values.Example branch:
{
"id": "source-question-id", "type": "Question.Choice", "options": [ { "label": "Yes", "branchTargetId": "target-question-id" } ] }
forms_validate returns visible question IDs, issues, and the exact serialized answer array. Questions skipped by branching are omitted.
Browser bootstrap exposes the anti-forgery token, server session ID, form-owner tenant and user, responder tenant and user, and form prefetch URL. The client combines that data with scoped Forms cookies for direct requests.
Observed endpoint families:
GET /formapi/api/{tenant}/users/{owner}/light/runtimeForms('{formId}')
POST /formapi/api/{tenant}/users/{owner}/forms('{formId}')/responses POST /formapi/api/{tenant}/users/{o
Only the beginning is shown; read the rest in the repo