Onboarding request: io.github.QVerisAI/mcp for GitHub MCP Registry

[FairyZhang-sunny]

Request

Please onboard io.github.QVerisAI/mcp to the GitHub MCP Registry (github.com/mcp) so it becomes discoverable in the VS Code MCP Gallery.

Per the process confirmed in discussion #1257, the server is already published to the Official MCP Registry; this request is for the remaining manual initial curation step. Once onboarded, future versions can sync from the Official Registry.

Server details

Field	Value
Official Registry name	io.github.QVerisAI/mcp
Latest version	0.14.5 (active + latest)
Repository	https://github.com/QVerisAI/qveris-agent-toolkit
Manifest	https://github.com/QVerisAI/qveris-agent-toolkit/blob/main/packages/mcp/server.json
Documentation	https://github.com/QVerisAI/qveris-agent-toolkit/blob/main/docs/en-US/mcp-server.md
License	MIT
Remote transport	Streamable HTTP — https://mcp.qveris.ai/mcp
Authentication	MCP OAuth discovery, OAuth 2.1 authorization code + PKCE (S256), Dynamic Client Registration
Local fallback	npm — @qverisai/mcp (stdio)

What QVeris provides

QVeris gives agents a compact discovery-and-execution layer for a large catalog of professional data sources and external APIs without registering every downstream capability as a separate MCP tool. The six canonical MCP tools are:

• discover — find relevant capabilities by natural language
• inspect — inspect schemas, examples, reliability, and billing metadata
• probe — validate parameters and quote without executing
• call — execute the selected capability
• usage_history — audit calls and outcomes
• credits_ledger — reconcile credit activity

The manifest includes explicit MCP tool annotations. Execution remains an explicit call step; discovery and inspection do not execute external capabilities.

Verification

Official Registry latest-version document:

https://registry.modelcontextprotocol.io/v0.1/servers/io.github.QVerisAI%2Fmcp/versions/latest

It reports version 0.14.5, status: active, isLatest: true, the hosted Streamable HTTP endpoint, logo metadata, and the npm fallback.

OAuth discovery endpoints are public and match the registered MCP resource:

• Protected resource metadata: https://mcp.qveris.ai/.well-known/oauth-protected-resource/mcp
• Authorization server metadata: https://auth.qveris.ai/.well-known/oauth-authorization-server

The authorization metadata advertises PKCE S256 and the Dynamic Client Registration endpoint. An unauthenticated MCP request returns 401 with a WWW-Authenticate: Bearer challenge pointing to the protected-resource metadata.

Current GitHub Registry query:

https://api.mcp.github.com/v0.1/servers?search=qveris

As of 2026-10-12, it returns total: 0.

Readiness

• Public, actively maintained MIT-licensed repository
• Hosted endpoint and npm fallback both represented in the Official Registry manifest
• Browser OAuth flow suitable for VS Code; no API key needs to be pasted into Gallery installation
• English and Chinese VS Code setup documentation
• Release workflow validates the Registry manifest and publishes through GitHub OIDC
• Registry preparation merged in feat(mcp): prepare hosted listing for VS Code gallery QVerisAI/qveris-agent-toolkit#333

Could the product/curation team please consider onboarding QVeris so Official Registry versions can sync into the GitHub MCP Registry and VS Code MCP Gallery? We are happy to provide any additional validation or metadata needed. Thank you.