MCP App Store
by KeenEthics
selfboo.st icon

selfboo.st MCP Server for ChatGPT

By T1M LLC

Overview

selfboo.st helps the people who build MCP servers see how AI agents choose between servers, and why theirs is passed over. Connected to ChatGPT, it works with the servers in your selfboo.st workspace, so you can check and improve your tools from a conversation. What you can do: - List the MCP servers in your workspace with their tool counts, latest audit score and latest Pick Score. - Read a server's audit: a 0–100 score across five dimensions, every failing check with its fix, and a suggested rewrite of each tool's title, description and parameters. - Fetch a server's tools again after you change them, and audit them again. - See the intents a server is tested on (the jobs it should win) with their prompts and competitors. - Start a pick test and read its result: how often AI models choose your server over its competitors for your prompts, with a 90% interval, by model, by intent and by competitor. Good to know: - Pick tests are simulations. They ask AI models, through their APIs, which connector they would suggest when shown your listing and your competitors'. They are not ChatGPT's or any other app's real ranking, and every score names the models it ran on. - selfboo.st reads a server's tools by sending it initialize and tools/list only. It never calls a server's tools and never changes your server: it suggests, and you decide what to ship. - Servers, intents, prompts and competitors are added in the selfboo.st app; from ChatGPT you read and test them. - You need a selfboo.st account.

How to connect

  1. ChatGPT: Open the selfboo.st listing in the ChatGPT app directory and choose Connect. ChatGPT listing selfboo.st exposes 6 tools here.

Tools

get_audit

ChatGPT
Returns the latest audit of one of your MCP servers: the 0–100 score, the five dimension scores, the failing checks with their fixes, and the model's suggested rewrite for each tool. Use this when the user asks why agents pass over their server or how to improve its tool descriptions, and after run_audit. An audit still running says so instead of a score; the rewrites follow the score by a few seconds. Not for running a new audit — use run_audit.

get_pick_test

ChatGPT
Returns a pick test's status and, once it is done, the Pick Score with its 90% interval, the score by model and by intent, and each competitor's share of the picks. Use this after run_pick_test, or with a server's pickTestId from list_servers when the user asks how its last test went. Not for starting a test — use run_pick_test. Every number is from a simulation, not the host's real ranking.

list_intents

ChatGPT
Lists the intents of one server — the customer jobs it should win — with their prompt and competitor counts and last Pick Score. Use this before run_pick_test to choose which intents to test. Not for creating intents; those are written in the selfboo.st app.

list_servers

ChatGPT
Lists the MCP servers in your selfboo.st workspace, each with its id, tool count, latest audit score, and latest Pick Score with the id of the pick test that measured it. Use this first, whenever the user asks about their servers or you need a server_id or test_id for another tool. Not for searching other people's servers.

run_audit

ChatGPT
Fetches a server's tools again and starts a new audit. Use this when the user has just changed their server's tool descriptions and wants a fresh score. The audit takes about thirty seconds; read it with get_audit. Not for reading the last audit — use get_audit. Charges a few credits for the model review, again on every call.

run_pick_test

ChatGPT
Starts a pick test: a simulation that asks AI models, prompt by prompt, which connectors they would suggest from your server and its competitors, and counts how often they pick yours. Use this when the user asks how often agents choose their server or wants to compare against competitors. Returns the test id and its credit cost; read the result with get_pick_test. Not for reading a finished test — use get_pick_test. Spends credits, and every call starts and pays for a new test.

Capabilities

Writes

Example Prompts

Click any prompt to copy it.

Similar Apps

Listed on MCP App Store

Listed on MCP App Store

Add this badge to your site or README. It links to this page.

HTML

<a href="https://mcp-app-store.keenethics-labs.com/app/selfboo-st"><img src="https://mcp-app-store.keenethics-labs.com/badges/listed-on-mcp-app-store.svg" alt="Listed on MCP App Store" width="220" height="44"></a>

Markdown

[![Listed on MCP App Store](https://mcp-app-store.keenethics-labs.com/badges/listed-on-mcp-app-store.svg)](https://mcp-app-store.keenethics-labs.com/app/selfboo-st)