Skip to main content

​
Feedback Triager

Feedback Triager watches a Slack channel you designate as your feedback inbox. Each fresh top-level message starts one session that claims the thread, subscribes to follow-ups, and classifies the report. The Expert stays focused on triage: it can read code, logs, and metrics to explain a problem, but it never changes code itself. For bugs, regressions, technical failures, and operational problems, Feedback Triager automatically investigates the root cause when the cause matters. It combines runtime and static evidence when both are available, labels uncertainty clearly, and stops when the evidence supports an actionable conclusion. It skips root-cause analysis for praise, feature requests, and questions with a direct documented answer.

​
Tickets and next steps

Feedback Triager can create tickets in Linear, Jira, or GitHub Issues. During setup, choose one of two modes:
  • Automatic: Create a ticket after classification and any applicable investigation, once the category-specific quality gate passes.
  • On request: Offer ticket creation when useful, but create it only after a clear request in the thread.
Before filing, the Expert searches the configured project, board, or repository for plausible duplicates. It does not create a ticket for every message: pure praise receives a reaction, direct questions are answered directly, and incomplete reports receive at most one focused clarification question for each gap. When action is useful, the reply presents one numbered Next Steps list. Depending on the findings, options can include creating a ticket, routing a concise handoff to another configured Slack channel, gathering specific missing evidence, or launching PR Author for a recommended fix. Code changes require a clear human request, and repeated requests reuse the existing result instead of repeating side effects. Feedback Triager also remembers explicit standing guidance for classification, routing, deduplication, and when to stay silent within each configured channel.
  • Best for: Teams with a busy feedback or bug-report channel who want evidence-backed triage and consistent follow-through without turning every conversation into a ticket.
  • How to use: This Expert is event-driven — you don’t launch sessions from it manually. Connect Slack and one supported ticket tracker, configure the feedback channel, ticket-creation mode, PR or MR Author worker, target repository, and any evidence sources or routing destinations. The Expert then handles each new top-level message in its own long-lived thread session.

​
See Also

  • PR Author — the worker the triager dispatches when a human requests a recommended fix.
  • Understanding Automation — how triggers and subscriptions drive event-driven Experts like this one.