mcp-serverSAFE
MCP server that sends your AI agents' tasks to iPhone, iPad, and Apple Watch
Overview
From the repository's own README, as read at the audited commit. Badges and raw HTML are left out.
[!NOTE] The new iOS app, Relaydeck, is on its way to the App Store. The download link goes here when it is live.
[](https://github.com/Aayush9029/mcp-server/actions/workflows/ci.yml)
Verified with
Task Management MCP Server
Bridge AI assistants with real-world task management
This MCP server connects Claude Code, Cursor, Codex, and other AI tools to your iPhone, iPad, and Apple Watch. Your agents create, update, and track tasks, and your devices show each change live.
What You Can Do
- Ask Claude to create tasks during coding sessions
- Let Cursor update task status as it completes work
- Get notifications on your iPhone when AI assistants finish tasks
- Follow work in progress with a Live Activity on the Lock Screen
- Answer your agent's questions from the notification
- Assign a task or send a message from your phone, and your agent picks it up
- Track progress across multiple AI tools and automation scripts
Perfec
c847ca0bace2OBSERVED · 2026-10-11Connect
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 mcp-client-server --env STUB_GOOD_KEYS=${STUB_GOOD_KEYS} -- uvx mcp-client-server{
"mcpServers": {
"mcp-client-server": {
"command": "uvx",
"args": [
"mcp-client-server"
],
"env": {
"STUB_GOOD_KEYS": "${STUB_GOOD_KEYS}"
}
}
}
}Trust audit
SAFEgrade B · trust 83/100 Nothing in the source contradicts what it says it does. Grade A is reserved for packages that have also passed the behavioural sandbox.
| 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 | PASS |
| L4 | Behavioural (sandbox) | SKIPPED |
What the source does
- Filesystem
- none-observed
- Network
- declared (6 observation(s))
- Shell
- none-observed
- Dependencies
- pinned
- Secrets in source
- none-found
Findings (25)
parser.add_argument("--url", default="http://127.0.0.1:8000")art-facets.avif
art-paper.avif
ipad-task-detail.avif
iphone-keys.avif
iphone-settings.avif
link.symlink_to("../../dotfiles/cursor-mcp.json")return f"http://127.0.0.1:{self.server_address[1]}"yield f"http://127.0.0.1:{port}"done = hook.run("claude", event, env={"RELAYDECK_URL": "http://127.0.0.1:9"})outputs.append(hook.run("claude", "start", env={"RELAYDECK_URL": "http://127.0.0.1:9"}))"attestation_object": "o2NmbXRvYXBwbGUtYXBwYXR0ZXN0Z2F0dFN0bXSiY3g1Y4JZBCEwggQdMIIDo6ADAgECAgYBnbE/C04wCgYIKoZIzj0EAwIwTzEjMCEGA1UEAwwaQXBwbGUgQXBwIEF0dGVzdGF0aW9uIENBIDExEzARBgNVBAoMCkFwcGxlIEluYy4xE
return base64.b64decode(value, validate=True)
attestation=base64.b64decode(SAMPLE["attestation_object"]),
key_id=base64.b64decode(SAMPLE["key_id"]),
assert hashlib.sha256(verified.public_key).digest() == base64.b64decode(SAMPLE["key_id"])
assert json.loads(base64.b64decode(query["config"][0], validate=True)) == {One Python service for the Relaydeck iOS app. AI agents reach it over MCP with an API key and report tasks. Apple devices reach it over REST and a server-sent event stream, and get pushes through APNs
- The MCP endpoint also accepts `?key=mck_<secret>`, for a client that cannot send headers. A URL can appear in logs, so use a header if you can.
<p class="card-caption">No email. No password. An open source server.</p>
<li><strong>Apple.</strong> To send a notification or to update a Live Activity, the server sends the push token and the content of the notification, for example a task title, a question, and a key la
<li>Open the app and make an API key for your agent.</li>
curl -fsSL https://relaydeck.onrender.com/install | sh
<code class="install-code" tabindex="0"><span class="install-prompt" aria-hidden="true">$ </span><span data-copy-source>curl -fsSL {{ base_url }}<wbr>/install | sh</span></code>Gates applied: no_behavioural_pass.
c847ca0bace2full audit observations/trust-audit/mcp-server/aayush9029__mcp-server.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 |
|---|---|---|---|---|---|
| 2026-10-11 | c847ca0bace2 | SAFE | B | 83 | first audit |
Questions
What is the mcp-server MCP server?
MCP server that sends your AI agents' tasks to iPhone, iPad, and Apple Watch
Is mcp-server safe to connect to an agent?
The audit found nothing in the source that contradicts what it says it does, and graded it B (83/100). Grade A is held back for packages that have also passed a sandboxed behavioural run, which is why a clean server reads B.
What credentials does mcp-server need?
It reads STUB_GOOD_KEYS 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 mcp-server run?
It speaks streamable-http, so it runs as a service you connect to over the network. It is published on PyPI as mcp-client-server.
How current is this page?
The grade is for one exact copy of the source (c847ca0bace2), read on 2026-10-11. The repository is watched and re-audited when it changes.