Cabinet Assistant MCP connector
Manages tasks and cases in Notion databases, including creating, assigning, editing, rescheduling, completing tasks, and adding, assigning, verifying cases with preview-confirm workflow.
Provides fast read access to patient and staff data, including listing daily tasks, team tasks, cases, and finding patients or staff members.
Click on "Install Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@Cabinet Assistant MCP connectorlist my tasks for today"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
Cabinet Assistant — MCP connector
A small, standalone, detachable remote-MCP service that gives Dr Amraoui's Claude a short menu of one-call actions over the cabinet's Notion (tasks + cases) and Supabase (fast reads). It never touches the cockpit web app — same databases, same field mappings, separate service.
Google Calendar stays on Claude's own Google Calendar connector; this service only returns the
event ids to create/update/delete, and stores them back on the task (link_event).
What it exposes (16 tools)
Reads (instant, from Supabase): list_my_day, list_team_tasks, list_cases, find_patient, find_staff
Task writes (preview → confirm): add_personal_task (private), add_clinic_task, edit_task,
reschedule_task, complete_task, assign_task, plus link_event (mechanical).
Cases (preview → confirm): add_case (+ auto triage task, web-app parity), assign_case, verify_case, set_report_status.
Every write returns a confirm_token first (nothing is saved); it only writes when called again with that token. Single-use token = confirmation enforced + idempotent.
Related MCP server: Notion MCP Server
Environment (.env.local, same values as the cockpit)
Var | Meaning |
| cockpit Notion integration (already shared with the DBs) |
| same Supabase project |
| long random string; the endpoint is |
| optional; else resolved from |
Run locally
npm install
npm start # http://localhost:8787/health
node test/smoke.mjs # end-to-end MCP test (handshake, read, preview→commit, guard, cleanup)
node test/smoke2.mjs # clinic delegation + case (dossier + triage) parityDeploy on Render (free)
Repo: https://github.com/Esubalew-kunta/MCP (the connector is at the repo root).
Render → New → Web Service → connect this GitHub repo.
Root Directory = (blank / repo root), Build
npm install, Startnpm start, Health check/health, Plan Free.Add the env vars above (values from the cockpit's
.env.local).Deploy. Your MCP URL is:
https://<service>.onrender.com/mcp/<MCP_SECRET>
Keep it awake (free tier sleeps after 15 min)
Point a free pinger at GET https://<service>.onrender.com/health every ~4 min:
cron-job.org, UptimeRobot, or an n8n Schedule → HTTP Request node.
Connect in her Claude
Claude → Settings → Connectors → Add custom connector → paste the full MCP URL (with the secret). Works in the mobile app too. Then a normal chat can create/assign/edit tasks & cases by voice.
Move to her VPS later
Deploy the same folder on the VPS (npm install && npm start behind HTTPS), repoint the connector URL.
Detach (remove entirely)
Remove the connector in her Claude settings.
Suspend/delete the Render service. → Cockpit, web app, Notion, calendar all unaffected. Nothing to clean up.
Security notes (pilot vs production)
Pilot auth = capability URL (secret path). Good for a private demo on an obscure URL.
Before real patients: put it behind OAuth, rotate
MCP_SECRET, restrict egress. No PHI is stored here.
This server cannot be installed
Maintenance
Resources
Unclaimed servers have limited discoverability.
Looking for Admin?
If you are the server author, to access and configure the admin panel.
Latest Blog Posts
- Your AI Chatbot Just Exposed Your CEO's Salary to an InternBy Om-Shree-0709 on .Agent IdentityMCP SecurityOAuth Delegation
- Why MCP Servers Need Execution Sandboxing (And Why Your Current Stack Isn't Enough)By Om-Shree-0709 on .Agentic AiPrompt InjectionWebAssembly
MCP directory API
We provide all the information about MCP servers via our MCP API.
curl -X GET 'https://glama.ai/api/mcp/v1/servers/Esubalew-kunta/MCP'
If you have feedback or need assistance with the MCP directory API, please join our Discord server