Skip to main content
Glama
Sign In
RE-Butler MCP Server

RE-Butler MCP Server

by AndritHashani
OverviewSchemaSourceRelated ServersDiscussionsAdmin
Python
Remote

RE-Butler MCP Server

Outbound MCP server exposing RE-Butler's knowledge and validation tools to neighbour teams and Azure AI Foundry agents.

Tools

Tool

Auth

Description

search_re_knowledge

RE.Read

Hybrid semantic search over LLM Wiki

list_knowledge_areas

RE.Read

List available knowledge domains

get_requirement_template

RE.Read

Get User Story / EARS / Gherkin templates

get_re_page

RE.Read

Fetch a specific wiki page

validate_user_story

RE.Read

INVEST/QUS/EARS validation

create_work_item

RE.Admin

Create ADO work item (HITL Gate C)

Related MCP server: Azure DevOps MCP Server

Setup

pip install -e ".[dev]"
cp .env.example .env  # fill in values
python -m re_butler_mcp.server

Tests

pytest

Smoke Test

# Against local server
python scripts/smoke_test_client.py

# Against deployed server
python scripts/smoke_test_client.py https://ca-genai4re-mcp-dev.<region>.azurecontainerapps.io/mcp

Infrastructure

cd infra
az deployment group create \
  --resource-group rg-genai4re-dev \
  --template-file main.bicep \
  --parameters containerImage=myacr.azurecr.io/re-butler-mcp:latest acrName=myacr

Auth

  • Callers must present a valid Entra ID Bearer JWT.

  • App Roles: RE.Read (all reads), RE.Admin (write tools).

  • Pre-authorize client app registrations in the server's app registration manifest.

  • Managed Identity handles all outbound calls (Search, OpenAI, ADO).

HITL Gate C

create_work_item always returns approval_pending: true. The calling workflow (Foundry / Durable Functions) must implement the approval step before the work item becomes active. Do not implement Gate C inside Connected Agent tools — keep it in the workflow layer.

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

    This server cannot be deployed

    One endpoint for every MCP toolRayrun runs every MCP server behind one URL and holds the credentials. Each call is decided per client, tool, and argument, then recorded.ray.runAd

    Resources

    Related MCP Connectors

    Related MCP Servers