Connector · Engineering
A BugSnag teammate in Slack, on SmartBear's remote MCP server.
Chickpea gives a channel a named AI teammate that holds its own BugSnag connection. It signs in with SmartBear and calls BugSnag's remote MCP server, and you tick the tools it may use before anything is saved. Ask it in Slack what is new since the release, and it answers in the thread.
What people ask it.
The teammate is @oncall, in #eng. Pick a scenario to watch it play.
| Error | Where | Events | Users |
|---|---|---|---|
| TimeoutError | checkout/retry.ts:118 | 1,284 | 431 |
| TypeError | checkout/session.ts:44 | 96 | 74 |
| RangeError | cart/price.ts:12 | 12 | 9 |
What is new since the release Read only
@oncall what errors are new in checkout-web since the 2.14.0 release
@oncall which open errors in production hit the most people this week
A read changes nothing in BugSnag, and a long answer comes back as a table in the thread. SmartBear documents a tool that lists a project's open errors with filters, and another that lists the filter fields a project has, so the teammate can find out what it is allowed to filter on before it asks. What it sees is what the signed-in BugSnag account sees.
The trace, and who it hit Read only
@oncall show me the stack trace on the checkout timeout and which devices it happens on
@oncall how many people hit the cart crash, and on which app versions
Another read. SmartBear says its error tool returns the latest occurrence with its stacktrace, user context, and device information, plus summaries by app version and OS, plus a link to the error on the dashboard. The teammate quotes what came back rather than paraphrasing an error it did not open.
A release's stability, or what is slow Read only
@oncall what is the stability score on the last three production releases
@oncall which operations got slower at p95 after Tuesday
Releases come back with their stability score and target, and a single release comes back with the builds under it. Performance is its own set of read tools: span groups sorted by duration percentiles, CPU, memory, frame rates, or HTTP error rates, then individual spans and whole traces underneath.
Mark one error fixed or ignored Applies at once
@oncall mark the checkout timeout as fixed, the rollback landed at 4:41
@oncall ignore the deprecated SDK warning in checkout-web, we are not fixing it this quarter
One error changes on the spot, if the person who connected BugSnag ticked the editing tools. Chickpea permits eight operations on an error: open, fix, ignore, discard, undiscard, snooze, and linking or unlinking an issue. Changing an error's severity is not supported.
Clear a batch after the fix ships Asks first
@oncall mark every checkout-web error fixed that stopped firing after 2.14.1
@oncall ignore the four Android crashes from this morning, they have the same cause
A change across several errors comes back as the list of what would change and waits for a yes. A write also needs the message you sent to name it, because thread history, retrieved content, and memory cannot authorize one on their own.
On a schedule Applies at once
@oncall every weekday at 9am Eastern, post the errors that are new in production since yesterday
@oncall every Monday at 8am Pacific, list last week's releases with their stability scores
A schedule saves on the spot. It needs an explicit time zone, runs under the BugSnag connection of the person who asked, and fails with a named reason rather than running if that connection is disconnected.
Connect it
Five steps, from Admin or from Slack.
You sign in through a page Chickpea serves, never in Slack. The link the teammate posts is good for one action, for one teammate, for 24 hours.
- In Chickpea Admin, open the teammate, then Connections, find BugSnag, and choose Connect. Or ask in Slack,
@oncall connect BugSnag, and follow the link it posts. - Choose who uses it. Team is one shared BugSnag account that everyone who can use the teammate acts through. Personal means each member signs in as themselves, and BugSnag records their actions as them.
- Choose Continue to sign in, then Sign in with SmartBear. SmartBear asks you to authorize the BugSnag MCP server, and there is no token to paste and no config file to edit.
- Chickpea connects to bugsnag.mcp.smartbear.com/mcp and records the tools the server answered with. BugSnag is a reviewed connector, so no tool is granted yet.
- Tick the tools the teammate may use. The investigation tools are suggested; the editing ones are ticked only if you want them. Save, then ask something.
@oncall what is new in checkout-web todayis a good first question, because it reads and cannot change anything.
A connection belongs to one teammate. Giving a second teammate BugSnag means a second sign-in with its own review, so the support teammate never inherits what the on-call teammate can reach.
There is no built-in Chickpea skill for BugSnag, and none is needed. The tools arrive from SmartBear's remote server. Chickpea connects before it saves anything, records what the server answered with, and exposes only the ones you ticked. SmartBear describes the set as error monitoring, performance tracing, and release management, and lists each tool and what it returns. SmartBear lists the BugSnag tools and what each one does.
BugSnagBeside BugSnag
What the same teammate usually holds too.
Each is its own connection, granted to the teammate alone. All 37 connectors.
The other error tracker, for the service that has not moved onto this one
LinearThe crash becomes a Linear issue with the stack trace already in it
incident.ioWhen the crash turns into an incident, the same teammate can say who is on call
AtlassianA Jira ticket for the fix, beside the error it came from
Questions people ask
Is there a BugSnag MCP server?
Yes. SmartBear hosts one at bugsnag.mcp.smartbear.com/mcp and calls it the BugSnag Remote MCP Server. SmartBear says it gives an AI assistant access to BugSnag error monitoring, performance tracing, and release management with no installation, no Node.js, and no API tokens. The same tools also ship in SmartBear's npm package for people who would rather run a server locally. Chickpea connects to the hosted one, so a teammate in Slack uses it with no editor and no config file anywhere.
How do I connect to the BugSnag MCP server?
For a coding client, follow SmartBear's own page, which gives the config for VS Code with GitHub Copilot, Cursor, Claude Desktop, and one command line for Claude Code. Each one points at the same URL and opens a browser for a SmartBear ID sign-in. For Slack there is nothing to install. Open the teammate in Chickpea Admin, choose BugSnag, sign in with SmartBear, then tick the tools it may use. Or ask "@oncall connect BugSnag" in the channel and follow the link it posts.
What is BugSnag used for?
Error monitoring for web and mobile applications, plus performance tracing and release management, which is how SmartBear describes the tools its MCP server exposes. In practice a team watches which errors are open in a project, which occurrence of an error carries the useful stack trace, how stable a release is against its target, and which operations are slow at p95. Those are the questions the teammate answers in the channel.
Does it use BugSnag's MCP server or the API?
The MCP server. BugSnag is a vendor MCP lane: the deployment connects to bugsnag.mcp.smartbear.com/mcp over streamable HTTP and signs in with OAuth, asking for the api scope. There is no BugSnag auth token anywhere in this connector, and the connection cannot reach any other address.
Which tools does the teammate get?
The ones you tick when you connect, and no others. Most MCP presets hand the teammate everything the server answered with. BugSnag does not. It is a reviewed connector, so the sign-in ends on a page listing the discovered tools, with the investigation ones suggested and the editing ones left off. Your choices become the maximum the teammate can reach. Reconnecting later can narrow that grant or drop a tool whose shape changed, and it can never widen one.
Can I limit it to one BugSnag project?
Not through this connection. SmartBear's local package takes a project API key that narrows every search to one project, but the remote server signs you in instead, so the tools can reach every project the BugSnag account you signed in with can see. What you narrow here is the tool list and the account. If one project is the point, connect a personal account that only has that project, or say the project in your message and check the answer names it back.
Can it change an error's severity, or delete anything?
No on severity. SmartBear's update tool can override an error's calculated severity, but that path needs the server to ask a follow-up question mid-call, which the Slack runtime does not do, so Chickpea permits eight operations on an error and no others: open, fix, ignore, discard, undiscard, snooze, link issue, and unlink issue. Discard is the heavy one: BugSnag keeps the events already in the dashboard and stops storing new ones for that error, and undiscard puts it back. Beyond that, a write needs your current message to name it, because thread history, retrieved content, and memory cannot authorize one on their own.
Do I still need BugSnag's Slack integration?
Yes, for anything that starts on BugSnag's side. BugSnag's own Slack app posts to a channel when an error first appears, when a fixed error comes back, when events spike, when a release goes out, and at milestones like the 100th and 1000th event. Those messages carry buttons that change the error's state. A Chickpea teammate hears nothing from BugSnag and posts nothing unprompted. It answers when you ask, reads across releases and traces, and holds BugSnag beside your other connected accounts in the channel where the question came up.
Which BugSnag plan does this need?
SmartBear's MCP pages name no plan requirement for the remote server. What the teammate can see and change is what the account you signed in with can see and change, because it acts through that sign-in. BugSnag does name one plan gate nearby: the advanced filters on its own Slack notifications, which use saved filtersets, are documented as Preferred and Enterprise. Chickpea is Apache 2.0 and runs on your own Cloudflare account, so there are no seats and no metering on this side.
Where does the OAuth token live?
On your deployment, outside the teammate record. It never enters the model's context, the tool arguments the model writes, or Slack. The model picks a connection by id and the credential goes in at the last step on the way out. Disconnect the account and the stored secret is removed, and any schedule that depended on it stops.
Try it
Give a channel a BugSnag teammate.
One deploy into your own Cloudflare account, connect Slack, sign in to BugSnag, ask. Apache 2.0, no seats, no metering.
Or run it on Node. The setup guide covers both.
Last checked
Facts on this page were checked against the product and BugSnag's own pages on September 17, 2026. This page is also available as markdown at /connectors/bugsnag.md.