{"slug": "mcp-prompts-in-claude-code-a-simple-guide-for-developers", "title": "MCP Prompts in Claude Code: A Simple Guide for Developers", "summary": "Anthropic's Claude Code now dynamically discovers and exposes prompts from connected Model Context Protocol (MCP) servers as slash commands, enabling developers to invoke reusable workflows such as incident response or deployment checklists directly in the coding assistant. The integration, which uses the MCP prompts/list and prompts/get protocol methods, distinguishes prompts as user-controlled workflows from tools that Claude can execute autonomously, and complements persistent project instructions stored in CLAUDE.md.", "body_md": "**MCP (Model Context Protocol)** has made it much easier to connect Claude Code with the tools and systems developers already use.\n\nA development team can connect Claude Code to GitHub, Jira, databases, monitoring platforms, CI/CD systems, or internal DevOps services through an MCP server. But MCP is not limited to tools that *do* something. An MCP server can also provide **prompts** — reusable workflows that developers can invoke when they need them.\n\nAn MCP server can expose reusable prompt templates. For example, imagine your company has a DevOps MCP server\n\nThis is the interesting part. MCP prompts exposed by connected MCP servers automatically become available as commands in Claude Code. They are dynamically discovered from connected servers.\n\nOnce an MCP server is connected to Claude Code, its prompts are **dynamically discovered**.\n\nYou can type / in Claude Code and see the available commands. MCP prompts appear using a naming pattern like\n\n**/mcp__servername__promptname**\n\nFor example, a DevOps server named devops with an incident_response prompt would appear as\n\n**/mcp__devops__incident_response**\n\nA deployment checklist could be\n\n**/mcp__devops__deploy_checklist**\n\nThis is built into the MCP integration in Claude Code — you don’t have to manually recreate every MCP prompt as a local command.\n\nImagine your team has an internal DevOps MCP server.\n\nIt exposes tools such as\n\ndeploy_application\n\ncheck_pipeline\n\nget_deployment_status\n\nfetch_logs\n\nIt also exposes prompts\n\ndeploy_checklist\n\nincident_response\n\nrelease_review\n\nThe tools give Claude the ability to interact with systems.\n\nThe prompts provide a **predefined workflow or set of instructions** for a particular task.\n\nFor example, incident_response could contain a workflow like\n\n1. Identify the affected application\n\n2. Check recent deployments\n\n3. Review application logs\n\n4. Check service health and metrics\n\n5. Identify possible root cause\n\n6. Recommend remediation\n\n7. Prepare an incident summary\n\nInstead of typing that workflow every time, the team can expose it through the MCP server.\n\nThere is a simple flow happening behind the scenes\n\nAt the protocol level, MCP clients can discover prompts with prompts/list and request a selected prompt with prompts/get. The resulting messages are then added to the conversation.\n\nThis is why an MCP prompt feels more like a **reusable workflow** than a normal API call.\n\nThis is probably the most important distinction to remember.\n\nA tool represents something Claude can execute through the connected system.\n\nFor example\n\nget_deployment_status() or fetch_logs()\n\nThese provide an actual capability to interact with an external system\n\nA prompt gives Claude a predefined way of approaching a task.\n\nincident_response\n\nmight say\n\n*Investigate the production incident by reviewing the deployment, logs, metrics, and recent changes before recommending remediation*\n\nClaude can then use the available tools to perform that investigation.\n\nMCP treats prompts as **user-controlled**, while tools are designed to be invoked by the model when appropriate.\n\nAnother common source of confusion is CLAUDE.md.\n\nCLAUDE.md is where you keep persistent instructions for your project.\n\n# Engineering Guidelines\n\n- Use Java 21\n\n- Follow our REST API conventions\n\n- Use JUnit 5\n\n- Run unit tests before creating a PR\n\n- Never commit secrets\n\nThese instructions establish the normal rules Claude should follow while working in the project.\n\nAn MCP prompt serves a different purpose.\n\nThink of it this way\n\n**CLAUDE.md** — Persistent project instructions\n\n**Claude Code Skill** — Reusable local workflow\n\n**MCP PromptReusable** — workflow provided by an MCP server\n\n**MCP Tool** — Execute an action against an external system\n\n**MCP Resource** — Provide external information/context\n\nClaude Code now uses **skills** for reusable workflows and custom commands, while MCP handles connections to external services and capabilities.\n\nMCP prompts aren’t necessarily fixed.\n\nA prompt can define arguments that the developer supplies when invoking it.\n\n**/mcp__github__pr_review 456**\n\nHere 456 could represent the pull request number.\n\nA DevOps workflow might look like\n\n**/mcp__devops__incident_response INC-4521**\n\nThe prompt can use that argument when generating the instructions for Claude. Claude Code supports passing prompt arguments after the MCP command, and MCP defines arguments as part of the prompt metadata.\n\nThis makes MCP prompts useful for real enterprise workflows rather than just simple static text.\n\nLet’s say your company has standardized its release process.\n\nThe DevOps MCP server provides\n\nPrompts\n\n**deploy_checklist\nrelease_review\nincident_response**\n\nA developer is preparing a deployment.\n\nThey run\n\nClaude receives the team’s standardized deployment workflow.It might then use MCP tools to\n\n→ Check CI/CD status\n\n→ Review deployment configuration\n\n→ Check recent changes\n\n→ Check application health\n\n→ Validate rollback readiness\n\nThe important point is that the **prompt defines the process**, while the **tools provide the capabilities** needed to carry it out.\n\nThis becomes particularly valuable when a team has repeatable processes.\n\nInstead of asking every developer to remember the company’s incident-management procedure, deployment checklist, or release-review process, those workflows can be made available through the MCP server.\n\n/mcp__devops__deploy_checklist\n\n/mcp__devops__incident_response\n\n/mcp__devops__release_review\n\nNow everyone starts from the same workflow. That can help teams standardize\n\n**Deployment → Incident response → Troubleshooting → Release validation → Security review**\n\nwhile still allowing Claude to use the appropriate tools and data sources.\n\nPrompt → Instructions / workflow\n\nResource → Information / context\n\nTool → Action / capability\n\nAnd within Claude Code\n\nMCP Server\n\n↓\n\nMCP Prompt\n\n↓\n\n/mcp__server__prompt\n\n↓\n\nClaude Code\n\nSo when you connect an MCP server, its prompts don’t sit hidden inside the server. Claude Code discovers them and makes them available as **commands** that you can explicitly invoke.\n\nMCP prompts are a simple but powerful way to bring **standardized workflows into Claude Code**.\n\nA team can keep its DevOps expertise in reusable prompts such as and expose them through an MCP server.\n\nDevelopers then invoke them directly /mcp__devops__incident_response Claude receives the workflow, and MCP tools can provide the actual access to systems needed to complete the task.\n\n**Prompt tells Claude how to approach the work. Tool gives Claude the ability to perform the work.**\n\nThat distinction is the key to understanding MCP prompts — and it’s also an important concept when working with Claude Code in real-world development environments.\n\n[MCP Prompts in Claude Code: A Simple Guide for Developers](https://pub.towardsai.net/mcp-prompts-in-claude-code-a-simple-guide-for-developers-2b7a8063f616) was originally published in [Towards AI](https://pub.towardsai.net) on Medium, where people are continuing the conversation by highlighting and responding to this story.", "url": "https://wpnews.pro/news/mcp-prompts-in-claude-code-a-simple-guide-for-developers", "canonical_source": "https://pub.towardsai.net/mcp-prompts-in-claude-code-a-simple-guide-for-developers-2b7a8063f616?source=rss----98111c9905da---4", "published_at": "2026-09-07 12:01:03+00:00", "updated_at": "2026-09-07 12:26:16.943346+00:00", "lang": "en", "topics": ["developer-tools", "ai-tools", "ai-agents"], "entities": ["Anthropic", "Claude Code", "Model Context Protocol"], "alternates": {"html": "https://wpnews.pro/news/mcp-prompts-in-claude-code-a-simple-guide-for-developers", "markdown": "https://wpnews.pro/news/mcp-prompts-in-claude-code-a-simple-guide-for-developers.md", "text": "https://wpnews.pro/news/mcp-prompts-in-claude-code-a-simple-guide-for-developers.txt", "jsonld": "https://wpnews.pro/news/mcp-prompts-in-claude-code-a-simple-guide-for-developers.jsonld"}}