| name | bugsnag |
| description | Interact with Bugsnag error monitoring via the SmartBear CLI. Use this skill whenever the user mentions Bugsnag, error monitoring, exceptions, stacktraces, error tracking, crash reports, or wants to investigate production errors. Also trigger when user asks about project errors, error events, or exception chains. Even if the user just says "check errors" or "what's breaking in production" — use this skill.
|
Bugsnag CLI
CLI tool for querying Bugsnag error data. Runs through MCPHub (must be running on localhost:9700).
smartbear <subcommand> [flags]
All commands support -o text|json|markdown|raw (default: text). Use -o json when you need to extract IDs for subsequent commands.
Interactive Workflow
Guide the user step-by-step through error investigation. At each step, present the results and let the user choose what to drill into next.
Step 1: Identify the project
Start by listing organizations, then projects. Present projects as a numbered list so the user can pick one.
smartbear smartbear-list-organizations -o json
smartbear smartbear-list-projects --organization-id ORG_ID -o json
Cache organization_id and project_id — they're reused in all subsequent calls.
If there's only one organization, skip the choice and proceed automatically.
Step 2: Show errors
List errors and present them to the user. Default to open errors, newest first, limit 10.
smartbear smartbear-list-errors \
--project-id PROJECT_ID --status open --sort newest --limit 10 -o json
Present as a summary table: error class, message, count, first/last seen. Ask the user which error to investigate, or let them search:
smartbear smartbear-search-issues \
--project-id PROJECT_ID --query "timeout" -o json
Step 3: Dive into the error
Once the user picks an error, show the latest event and stacktrace together:
smartbear smartbear-view-latest-event \
--error-id ERROR_ID -o json
smartbear smartbear-view-stacktrace \
--project-id PROJECT_ID --event-id EVENT_ID --include-code true -o json
Present a clear summary: what happened, where in code, how often.
Then offer next steps:
- Exception chain — if the error might wrap a root cause
- Full tabs — for complete context (app, device, user, request, breadcrumbs)
- Other events — to see if the error manifests differently across occurrences
- Back to error list — to look at other errors
Step 4: Deep-dive (on user's choice)
smartbear smartbear-view-exception-chain \
--project-id PROJECT_ID --event-id EVENT_ID -o json
smartbear smartbear-view-tabs \
--project-id PROJECT_ID --event-id EVENT_ID --include-code true -o json
smartbear smartbear-list-error-events \
--project-id PROJECT_ID --error-id ERROR_ID --limit 5 -o json
smartbear smartbear-view-event \
--project-id PROJECT_ID --event-id EVENT_ID -o json
Commands Reference
Discovery
| Command | Flags | Returns |
|---|
smartbear-list-organizations | none | org list with IDs |
smartbear-list-projects | --organization-id ID | projects in org |
Error Investigation
| Command | Required | Optional |
|---|
smartbear-list-errors | --project-id | --status open|fixed|ignored --sort newest|oldest|priority --limit N |
smartbear-view-error | --error-id | — |
smartbear-search-issues | --project-id | --query TEXT --error-class CLASS --app-version VER |
Event Deep-Dive
| Command | Required | Optional |
|---|
smartbear-list-error-events | --project-id --error-id | --limit N |
smartbear-view-latest-event | --error-id | — |
smartbear-view-event | --project-id --event-id | — |
smartbear-view-stacktrace | --project-id --event-id | --include-code true |
smartbear-view-exception-chain | --project-id --event-id | — |
smartbear-view-tabs | --project-id --event-id | --include-code true |
Tips
view-tabs is the most comprehensive — use when user needs full context
--include-code true shows source code around each stack frame (if available)
- MCPHub must be running:
make status in /home/danil/code/mcp-hub
- When presenting errors to user, summarize key fields (class, message, count) — don't dump raw JSON