# FireHydrant MCP for AI Agents AI Agent Connect > FireHydrant MCP lets you manage incidents, services, and responder teams through your AI client. It connects your FireHydrant account to tools like Claude or Cursor so you can declare outages, update timelines, and check runbooks using natural language. It's built for high-pressure situations where speed and accuracy matter most. ## Overview - **Category:** devops-cicd - **Price:** Free - **Endpoint:** https://edge.vinkius.com/vk_preview_Z9fBTFG9bYGC52Uaccytuw08nwfsGKdNijiOjVX5/ai-agent-connect - **Tags:** incident-management, site-reliability, service-catalog, incident-response, on-call, post-mortem ## Description Imagine an outage happens at 2 a.m. instead of digging through multiple tabs to find the right team or checking a messy wiki for a runbook, you just tell your AI client what's happening. You can declare the incident, pull up the service catalog to see what's down, and identify the right responders in seconds. It handles the tedious parts of incident management, like posting notes to a timeline or looking up recent change events that might have caused the spike. This Connector makes it easy to keep everyone on the same page without leaving your primary workspace. By connecting to your account through Vinkius, you turn your AI into a coordinated incident commander that knows your infrastructure. You get to focus on fixing the problem while the agent handles the data gathering and status updates. ## Tools ### add_incident_note Add a note to an incident timeline. This helps keep responders informed of the latest investigation steps. ### create_incident Create a new incident in FireHydrant. Use this to quickly declare an outage and notify the team. ### list_services List the service catalog. Use this to see all defined services and their relationships. ### list_teams List responder teams. This helps you find the right people to assign to a new incident. ### update_incident Update an existing incident. Use this to change milestones or status fields as the situation evolves. ### get_incident Get specific details for an incident. This helps you see the status and severity of a particular issue. ### get_service Get details about a specific service. Use this to understand dependencies and impact during an outage. ### get_team Get details about a responder team. This helps you identify who is on call and available to help. ### list_change_events List recent change events. Use this to see if a recent deployment or config change caused the issue. ### list_incidents List all incidents. This gives you a high level view of every active issue in your environment. ### list_retrospectives List retrospectives. Use this to find previous post-incident reviews for learning. ### list_runbooks List active runbooks. This helps you find the correct automated workflows for a specific problem. ## Prompt Examples **Prompt:** ``` Show me all active incidents right now. ``` **Response:** ``` ### Active Incidents I've pulled the current list from FireHydrant: | Incident ID | Name | Severity | Status | | :--- | :--- | :--- | :--- | | `inc_442` | **Payment Gateway Timeout** | Sev-1 | Active | | `inc_509` | **Slow DB Queries** | Sev-2 | Investigating | Would you like me to get more details on either of these? ``` **Prompt:** ``` I need to declare a new sev-1 incident for the checkout page being down. ``` **Response:** ``` I've declared that incident for you. **New Incident Created** - **Name:** Checkout Page Down - **Severity:** Sev-1 - **ID:** `inc_667` I've also notified the responder teams. Do you want me to add an initial note to the timeline? ``` **Prompt:** ``` What runbooks do we have for the database connection issues? ``` **Response:** ``` I found the following active runbooks related to database issues: * **DB_CONN_01**: Standard connection pool reset * **DB_CONN_02**: Failover to secondary region * **DB_CONN_03**: Manual index rebuild Which one would you like to see the full steps for? ``` ## Capabilities ### Declare new incidents Create a new incident record instantly to notify your team and start the response. ### List active incidents See every ongoing outage in your environment at a glance. ### Update incident milestones Change status fields and milestones as the situation evolves. ### Check service catalog details Access service definitions to understand dependencies and impact. ### Identify responder teams Find the correct people to assign to an incident based on your team list. ### Post timeline updates Add notes directly to an incident timeline to keep responders informed. ### View active runbooks Retrieve the correct automated workflows for specific technical problems. ### Retrieve retrospectives Pull up previous post-incident reviews to identify patterns and learn. ### Track infrastructure changes List recent change events to see if a deployment or config update caused a spike. ## Use Cases ### Outage Response An SRE notices a database spike. They ask the agent to list_incidents, find the right one, and then use create_incident to declare a new sev-1. ### Runbook Retrieval During a web server failure, an engineer asks the agent to list_runbooks to find the correct recovery steps for the load balancer. ### Stakeholder Updates An engineering manager asks the agent to list_change_events to see if a deployment caused the current latency before replying to a status page. ### Team Coordination An incident commander uses list_teams to find the on-call rotation for the networking team to bring them into a live incident. ## Benefits - Faster incident declaration using create_incident means you spend less time filling out forms and more time fixing the site. - Real-time timeline updates with add_incident_note keep everyone informed without manual copy-pasting between tools. - Quick access to runbooks via list_runbooks ensures you follow the correct procedures during high-pressure outages. - Instant service impact analysis with list_services helps you identify which dependencies are failing first. - Better root cause analysis by using list_change_events to see what changed in the infrastructure right before a spike. - Easier post-mortem preparation by using list_retrospectives to gather past data for your next review. ## How It Works The bottom line is it turns your AI into a direct interface for your entire incident management platform. 1. Subscribe to the FireHydrant MCP on the Vinkius Marketplace. 2. Add your FireHydrant API Key from your account settings. 3. Use your AI client to manage incidents and services. ## Frequently Asked Questions **Does the FireHydrant MCP work with Claude?** Yes, it works with any MCP-compatible client, including Claude, Cursor, and Windsurf. **How do I add my FireHydrant API key?** You can find your API Key in your FireHydrant account settings under the API Keys section. **Can my agent update the incident timeline for me?** Yes, your agent can post notes directly to the timeline, which keeps your entire team updated in real time. **Can I see my service catalog with this Connector?** Yes, your agent can list all services in your catalog to help you understand dependencies during an outage. **Is this Connector built for SRE teams?** It is specifically designed for SREs and DevOps engineers who need fast, reliable access to incident data. **Can it help my team with post-mortems?** Yes, it can retrieve past retrospectives so you can quickly see how similar issues were handled in the past. **How do I declare a new incident using the agent?** You can use the 'create_incident' tool. Simply provide a name and optionally a summary and severity level. The agent will trigger the declaration in FireHydrant and return the new incident details. **Can I see recent infrastructure changes using this server?** Yes! The 'list_change_events' tool retrieves a history of recent changes. This is extremely useful for identifying potential root causes during an active incident investigation. **How do I add a status update to an ongoing incident?** Use the 'add_incident_note' tool. You'll need the Incident ID and the text of your update. The note will be immediately posted to the incident's timeline for all responders to see.