skills.rest
denniszielke
denniszielkeCOMMUNITY
Explore all skills by this author →
WebsiteGitHub

ops

Orchestrates deployment, provisioning, and cleanup of Azure banking agents and MCP servers.

4|Updated Jul 1, 2026
Search Tags:#deployment#azure#MCP servers#runbook#azd#container apps#foundry agents
One-click install
npx skills add https://github.com/denniszielke/agentic-banking-ecosystem --skill ops-denniszielke
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: ops
Source: https://github.com/denniszielke/agentic-banking-ecosystem/tree/main/.github/skills/ops
Command: npx skills add https://github.com/denniszielke/agentic-banking-ecosystem --skill ops-denniszielke

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Deploying a multi-organisation agentic banking demo involves many ordered steps — provisioning Azure infrastructure, generating demo data, building container images, deploying MCP servers, registering Foundry toolboxes, creating search indexes, and deploying agents — and doing them out of order or with wrong configuration causes failures. This Skill provides the complete operational runbook for the agentic-banking-ecosystem repository. ## Core Features & Use Cases - Full Lifecycle Coverage: Guides every stage from azd up provisioning through data generation, ACR image builds, MCP server deployment, toolbox registration, search index ingestion, agent deployment, and cleanup. - Entra ID Auth Configuration: Documents how the MCP servers validate Entra JWT tokens in-app, how to grant agent identities the Mcp.Invoke role, and how to handle Conditional Access exclusions and WorkIQ OAuth setup. - Troubleshooting Gotchas: Captures hard-won operational knowledge such as agent identity rotation after redeploys, the paused Fabric capacity error 42131, and Agent 365 observability role grants. - Use Case: You need to stand up the entire Bank North / Bank South demo from scratch. Follow the runbook to provision infrastructure, generate seeded customer data, build and deploy the customer/product MCP servers with auth, register toolboxes, ingest knowledge into AI Search, and deploy all three agents. ## Quick Start Use the ops skill to deploy the full agentic banking stack starting with azd up and then deploying the MCP servers and agents in order.

Frequently Asked Questions about ops

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I deploy the agentic banking demo to Azure?▼

Run azd up to provision infrastructure and write the .env file, then generate demo data, build the MCP server images in ACR, deploy the MCP servers with --build --register, create and ingest the search indexes, and finally deploy the agents with the provided scripts.

How do I deploy the MCP servers as Container Apps?▼

Run python -m scripts.deploy_customer_data_mcp_server --build --register and the equivalent product data script from the repo root. The --build flag builds the image in ACR first, and --register publishes the server as a Foundry toolbox in the same run.

Why do agentic-identity MCP tool calls fail with status 400?▼

Agent redeploys rotate the hosted agent's Entra Agent Identity, so the new identity lacks the Mcp.Invoke role and is not excluded from the High Risk Agents Conditional Access policy. Re-grant the role and add the new identity object id to the policy's excludeAgentIdServicePrincipals.

Can I run the MCP servers and agents locally for development?▼

Yes, each server runs as a Python module from the repo root, for example python -m src.customer_data_mcp_server.server on port 8092. Set CUSTOMER_MCP_URL and PRODUCT_MCP_URL to the local endpoints to bypass the Foundry toolboxes for local agent development.

What does error 42131 'This SQL database has been disabled' mean?▼

This error means the backing Fabric capacity fabricbanking is paused, taking all its SQL databases offline, and is not an MCP or auth bug. Check the capacity state with az rest against the Microsoft.Fabric provider and resume the capacity before further debugging.

How do I clean up all deployed Azure resources?▼

Run python -m scripts.delete_agents for hosted agents, delete_container_apps for Container Apps, and delete_search_indexes for the search indexes. Finally run azd down to tear down all remaining Azure resources.

Related Skills

Explore More Software Engineering Skills→
pascalvanderheidenpascalvanderheiden

azd-deployment

Automate Azure deployments and infrastructure provisioning with azd and Bicep.

Community
Advanced
OpenLAIROpenLAIR
1.0k

phoenix-observability

Trace, evaluate, and monitor LLM applications with OpenTelemetry and self-hosted storage.

Official
Advanced
rrbandarrbanda

azure-search-documents-py

Integrate Azure AI Search vector, semantic, and hybrid search into Python applications.

Community
Advanced
merceralex397-collabmerceralex397-collab

azure-deploy

Execute Azure deployments for prepared applications using azd, Terraform, Bicep, and Azure CLI.

Community
Advanced