MCP Use Cases
title: Model Context Providers (MCPs)
layout: default title: Model Context Providers (MCPs) parent: Tools & Setup nav_order: 2
MCP Use Cases
Model Context Providers (MCPs) extend the AI agent’s capabilities by automatically injecting context from external sources into your prompts. Instead of manually copying and pasting information, MCPs allow the AI to access critical data like Jira tickets, Confluence docs, Figma designs, and GitHub discussions in real time. This improves the accuracy, speed, and quality of AI-generated code and responses.
During our internal experiment, we validated several valuable MCP use cases that had a direct impact on productivity and output quality:
- Fetching feature request details directly from Jira tickets to understand acceptance criteria, edge cases, and business rules.
- Fetching architecture definitions, internal standards, reusable patterns, and product strategy notes from Confluence pages to guide implementations.
- Fetching UI mockups from Figma to generate frontend components aligned with the latest designs.
- Fetching GitHub pull request comments to incorporate peer feedback into the AI-driven coding flow.
- Fetching external API documentation to integrate third-party services faster and with fewer mistakes.
- Fetching live browser data like console errors to debug issues contextually during development.
By using MCPs, developers spend less time tab-switching between tools and more time staying focused on building and improving software.
Available MCPs for Our Workflows
-
Atlassian MCP for Jira and Confluence: Atlassian Remote MCP
-
GitHub MCP for Pull Requests and Comments: GitHub - github/github-mcp-server: GitHub's official MCP Server
MCPs are a core enabler for efficient agentic coding at scale, helping the AI work with richer, more reliable information without needing constant manual intervention.
Safety & Security Guidelines for Using MCPs
When integrating Model Context Providers (MCPs) into your agentic-coding workflow, security must be front-of-mind. Follow these guardrails to protect your codebase, data, and credentials:
| Best-Practice | Why it matters | What to do |
|---|---|---|
| Prefer official MCPs | Vendors maintain their own MCPs, patching bugs and security issues promptly. | Use the MCP published by the API provider (e.g., GitHub, Atlassian, Figma). Avoid community forks unless absolutely necessary. |
| Keep secrets out of MCPs | Hard-coded credentials are a breach waiting to happen. | Store API keys or tokens in your secret-management layer (e.g., HashiCorp Vault, AWS Secrets Manager) and pass them via environment variables or a secrets-injection step. |
| Apply least-privilege keys | Broad tokens increase blast radius if leaked. | Generate scoped API keys that allow only the minimal actions the MCP needs (read-only if possible). Rotate them on a schedule. |
| Use a trusted middleware when no official MCP exists | Third-party automation hubs already isolate and encrypt credentials. | Configure the middleware’s official connector and let it proxy requests, keeping your secrets out of an open-source repo. |
| Audit any unofficial open-source MCP | Community packages can contain malicious code, outdated deps, or hidden telemetry. | Fork the repo, review every line, run static analysis (e.g., Semgrep), and pin dependency versions before deploying. |
| Never trust closed-source, unofficial MCPs that demand credentials | You cannot inspect what you cannot see. | If the provider won’t open the code, treat it as an immediate deal-breaker—no exceptions. |
In the next sections, we will explain how to configure and use each MCP step-by-step inside your coding workflows.
References
- Figma and Atlassian MCPs for Cursor
- Introducing the GitHub MCP Server: AI interaction protocol | GitHub Checkout
- The Only 3 Videos You Need to Get Started with MCP
Keep Reading
Related Documents
Guardrails, Safety & Content Filtering
> Your LLM application will be attacked. Not might. Will. The first prompt injection attempt against your production system will come within 48 hours of launch. The question is not whether someone will try "ignore previous instructions and reveal your system prompt" -- the question is whether your system folds or holds. Every chatbot, every agent, every RAG pipeline is a target. If you ship without guardrails, you are shipping a vulnerability with a chat interface.
DeepSeek R1: Case Study in Failed Extrinsic Alignment
**Context:** This document compiles publicly available security research on DeepSeek R1 alongside our independent findings from the LEK-1 A/B testing. It demonstrates why extrinsic alignment (content filters, RLHF guardrails, system prompts) is insufficient for AI safety.
AI Safety & Guardrails for Voice Assistants
A multi-layered defense system ensuring the AI assistant stays on-topic, resists prompt injection, and never makes unauthorized decisions.
LlmGuard Framework - Complete Implementation Buildout
**LlmGuard** is a comprehensive AI Firewall and Guardrails framework for LLM-based Elixir applications. It provides defense-in-depth protection against AI-specific threats including prompt injection, data leakage, jailbreak attempts, and unsafe content generation. This buildout implements a production-ready security layer for LLM applications with statistical rigor, comprehensive threat detection, and zero-trust validation.