Rootly
Rootly is an AI-native incident management platform that automates workflows and collaboration, integrating with Slack, PagerDuty, and other tools to streamline incident response.
Verdict
Common use cases
- Review open action items during standup
- Check follow-up tasks after an incident
- Delete completed action items from chat
- Pull details on a specific task by ID
- Track post-mortem work without leaving Slack
Integration
- Vendor
- Rootly
- Category
- other
- Auth
- API_KEY
- Tools
- 3
- Composio slug
rootly
Tools
- Delete Action Itemdestructive
This tool allows for the deletion of a specific action item in rootly. it complements the existing rootly list action items functionality by providing the ability to remove individual action items from the system.
- Get Action Item Details
This tool retrieves a single action item by its id from rootly using a get request to https://api.rootly.com/v1/action items/{id}. it provides detailed information about an action item including summary, description, kind, assigned to, prio
- List Action Items
This tool retrieves a list of all action items for an organization in rootly. action items are tasks or follow-up items that need to be completed during or after an incident, helping to track and manage incident-related tasks effectively.
Setup
Setup guide
- 11. In Switchy, open your workspace settings and navigate to the Integrations page. 2. Find Rootly in the MCP directory and click Connect. 3. You'll be prompted to enter a Rootly API key — generate one by logging into your Rootly dashboard, going to Settings > API Keys, and creating a new key with read and write permissions for action items. 4. Paste the key into Switchy and click Authorize. 5. Once connected, open any Space and type '@Rootly list action items' to test the connection. 6. If the MCP returns a list of your organization's action items, the setup is complete. 7. To invoke Rootly in any conversation, @mention it followed by your request (e.g. '@Rootly get details for action item 12345').
What teammates see: by default, memories from Rootly are scoped to the Space (PROJECT visibility) - you can mark any memory PRIVATE or share it ORG-wide.
Works well with
Top models
Compatibility data appears once enough Spaces have used this MCP together with a given model.
How Switchy teams use it
Starter prompts
List All Action Items
@Rootly list all action items for our organizationOpen in a Space →
Get Item Details
@Rootly get details for action item with ID 98765Open in a Space →
Delete Completed Item
@Rootly delete action item 54321 since we finished that taskOpen in a Space →
Check Post-Incident Tasks
@Rootly show me action items from the last 48 hoursOpen in a Space →
Summarize Open Work
@Rootly list all action items and group them by statusOpen in a Space →
Example outputs
Illustrative - representative of the model's voice and quality, not literal recordings.
@Rootly show me all open action items from incidents in the last 30 days, grouped by assignee
Here's a summary of open action items from recent incidents: **Alice Chen** has 4 items (2 from the database outage on Jan 15, 2 from API latency spike on Jan 22). **Bob Martinez** has 3 items related to the payment gateway incident on Jan 18. **Unassigned** shows 2 items from the CDN failure on Jan 10. Most items involve post-incident documentation or configuration changes. The oldest unresolved item is 18 days old.
This example shows how the MCP lists action items and the AI synthesizes them into a readable summary. Requires read scope on action_items. The MCP returns raw data; grouping and age calculations happen in the AI layer, so results depend on what fields Rootly exposes in its API response.
@Rootly create an action item for the current incident: update runbook with new rollback procedure, assign to platform-team
I've created action item #847 in Rootly: "Update runbook with new rollback procedure." It's assigned to platform-team with a due date of 7 days from now (default). The item is linked to incident INC-2024-0156 and marked as high priority. You can view it at https://rootly.com/action_items/847 or update details directly in Rootly's dashboard.
This demonstrates write capability—creating action items during an active incident. Requires write scope on action_items. The MCP doesn't validate that 'platform-team' exists in your Rootly org, so typos will fail silently or create orphaned assignments. Always verify the assignee format your Rootly instance expects.
@Rootly get details on action item #823 and tell me if it's blocking any other work
Action item #823: "Migrate logging pipeline to new Datadog instance." Created 12 days ago, assigned to DevOps, currently **overdue by 5 days**. Status shows 'in progress' with a note added yesterday: 'Waiting on security review.' This item is tagged as a dependency for incident INC-2024-0142's resolution. Based on the context, it's likely blocking final closure of that incident and any related capacity planning work.
Shows how fetching a single action item's details enables the AI to reason about project dependencies. Requires read scope. The 'blocking' inference comes from the AI interpreting tags and notes—Rootly's API may not expose formal dependency graphs, so treat this as contextual analysis, not a guaranteed relationship map.
Use-case deep-dives
When Rootly MCP beats spreadsheets for follow-up tasks
A 6-person SRE team runs a weekly incident review where they assign action items from the last sprint's outages. The Rootly MCP is the right call here because it pulls the full list of open action items directly into Switchy, lets the team update statuses in chat, and deletes completed tasks without opening the Rootly web UI. The three tools cover the core loop: list items to see what's pending, get details to confirm ownership and due dates, and delete when work ships. This breaks down if your team creates more than 30 action items per week—at that scale you need Rootly's native filtering and bulk ops, not a chat interface. If your post-mortems generate 5-15 follow-ups per incident and you want one shared view during standups, this MCP keeps the team in Switchy instead of context-switching to Rootly.
Why this MCP works for support-to-engineering escalations
A 4-person customer success team escalates incidents to engineering and needs to confirm that follow-up tasks are logged and assigned. The Rootly MCP fits this workflow because CS can query action items by incident ID in Switchy, verify the engineer's name and summary, and close the loop without asking for a Rootly login. The get-details tool shows who owns the task and when it's due, which is the exact handoff data CS needs to update the customer. This setup assumes your CS team doesn't create or edit action items—they only read and confirm. If CS needs to assign tasks or change priorities, the MCP's three tools won't cover it and you'll need full Rootly access. For read-only incident follow-up visibility between support and engineering, this MCP delivers the handoff data in the team's shared workspace.
When Rootly MCP streamlines retro agenda building
A 5-person platform team holds a Friday retro to review the week's on-call incidents and their outstanding action items. The Rootly MCP is the right tool because it pulls the full action item list into Switchy before the meeting starts, so the team can triage which tasks to discuss without opening Rootly tabs. The list tool shows all pending items, the details tool surfaces context for ambiguous summaries, and the delete tool lets the team mark quick wins as done during the retro itself. This works best when your team generates 10-20 action items per week and wants to batch-review them in a single session. If your retros run longer than 30 minutes or involve editing task descriptions and reassigning owners, the MCP's read-and-delete scope is too narrow. For teams that want a fast weekly sweep of incident follow-ups in their shared chat workspace, this MCP keeps the retro moving.
Frequently asked
What does the Rootly MCP do in Switchy?
It lets your team view, create, and delete action items from Rootly incidents without leaving Switchy. The MCP pulls task lists from your Rootly organization and surfaces them in AI conversations, so you can track follow-ups or close completed items inline. It's useful if you run post-mortems in Rootly but want to reference or update action items while working in Switchy.
Do I need admin access to connect Rootly?
You need a Rootly API key with permission to read and write action items. Rootly generates API keys under Organization Settings → API Tokens. If your Rootly account restricts token creation to admins, ask an admin to create one for you. The key authenticates all requests, so treat it like a password—store it in Switchy's credential vault, not in shared docs.
Can the Rootly MCP create or update incidents?
No. This MCP only manages action items—the follow-up tasks tied to incidents. It cannot create, close, or modify incidents themselves. If you need to declare or update an incident, use Rootly's Slack bot or web UI. The MCP is narrowly scoped to post-incident task tracking, not the incident lifecycle.
Why use this instead of Rootly's Slack integration?
Rootly's Slack bot is better for real-time incident response—declaring incidents, paging on-call, posting updates. The Switchy MCP is better for async work: reviewing action items across multiple incidents, bulk-closing tasks, or referencing Rootly data in AI-generated reports. Use both. The MCP doesn't replace Slack; it gives you a query interface for action-item hygiene.
Who on the team should connect the Rootly MCP?
Whoever owns post-incident follow-through—usually an engineering manager, SRE lead, or incident commander. They'll generate the API key and add it to Switchy. After that, anyone in your Switchy workspace can invoke the MCP to pull action items. The connection is workspace-wide, so you only configure it once.