Tips and prompts for developers using the Cortex MCP

AI coding assistants are already transforming how developers work, helping them write code faster, answer tough questions, and automate repetitive tasks. It’s exciting, it’s powerful… and it’s just the beginning. 

Cortex MCP connects your AI assistant directly to your live service data, ownership, and organizational standards so it can give accurate, context-rich answers right in your IDE. With full organizational awareness of your architecture, dependencies, incident timelines, priorities, standards, and team ownership, AI assistants are becoming even more powerful.

→ Read the full MCP announcement

Adding Cortex in your IDE 

Installing the Cortex MCP is quick, simple, and requires no special configuration. In just a few steps, you can connect your AI assistant to your organization’s real-time service data and start asking smarter, context-aware questions right from your IDE.

Here’s a short video walking through the setup in less than 4 minutes. 

A few quick tips

• Be specific – The more context you give in your prompt, the faster you’ll get the right answer.

• Verify information – Ask your AI assistant to “only share data from Cortex” and “verify before answering” to keep it accurate.

• Chain your prompts – Follow up for deeper context or to explore recommendations.

• Ask for recommendations – Go beyond “what’s the current state?” and request next steps or scorecard-based guidance.

Top AI prompts for developers 

These categories cover the most common and high-value use cases we see across early Cortex MCP users.

1. Ownership and responsibility

Quickly find the right person or team to talk to. Cortex is the only IDP that fully solves service ownership—learn more in this blog.

• Who owns <service-name>?

• Show me all services owned by <team-name>.

• Who is on call for <service-name> right now?

• Which teams are responsible for APIs related to <feature>?

2. Service and system context

Get the full picture before making changes or troubleshooting.

• Summarize <service-name> and its dependencies, including any recent deploys.

• List services that depend on <service-name>.

• Show me the tech stack for <service-name>.

3. Scorecards and standards

Quickly assess progress against organizational standards.

• Give me a list of all of my services that aren’t currently at gold status.

• Which of my services need the most work to be compliant with all Scorecard standards?

• Give me recommendations on what to improve about <service-name> based on my organization’s best practices and active Initiatives.

4. Readiness checks

If you have production readiness scorecards setup, confirm if your services are ready to handle a launch or spike in traffic.

• We’re anticipating a lot of traffic next week. Are my services ready?

• Is <service-name> ready to ship? What’s outstanding?

5. Incident response

Find exactly what you need when things are on fire.

• Give me a timeline leading up to <incident-id>, including incident updates, recent deploys, dependencies, and on-call.

• Summarize info I need to handle an incident with <service-name>.

• What services are impacted by <incident-id>?

6. Migration readiness

If you have migration scorecards in place, access that data to make sure you’re ready for large-scale changes.

• What services owned by <person> still have pending tasks for the Helm upgrade scorecard?

• Which Tier 1 services have not yet been migrated to <new-platform>?

• List all services still using <deprecated-technology> and their owners.

7. Initiative tracking

Check progress and find blockers on time-bound work.

• What are my open initiatives in Cortex and what is left to complete? When are they due?

• Do I have any overdue initiatives?

• Show progress on the “Improve logging” initiative.

• Which services haven’t completed the API deprecation initiative?

• Who owns the last un-migrated services in this initiative?

8. Cross-team discovery

See how the organization fits together and locate the right services fast.

• Which services are owned by the data platform team?

• List all services tagged as ‘customer-facing’.

• What domains have the most services?

Developers can do more than ever with AI 

Whether you’re diagnosing a production issue, preparing for a big release, or onboarding to a new codebase, the right Cortex MCP prompt gets you the contextual information you need, right in your IDE. 

It’s just the beginning. Cortex is your partner for the full journey, from connecting AI to your engineering data today, to driving smarter, faster decisions across your organization tomorrow.

If you’ve already installed the Cortex MCP, share this list with your engineering team to get more value out of your AI. And if you’re not using it yet, all you need is 4 minutes to set it up and your AI assistant will instantly be able to surface the exact service owner, incident history, or readiness status you need without breaking your flow.

Not yet using Cortex? If maximizing your AI investments is a focus, book a demo and leave a note that you’re interested in the Cortex MCP server to help developers access everything they need directly from their IDE.