Turn Re:amaze bug reports into Linear issues, not duplicates
Every two hours we read your support inbox, spot the genuine product bugs, and file or update one Linear issue per problem.
Every two hours, review my recent Re:amaze support conversations, work out which ones describe a genuine product bug, and get those in front of engineering in Linear without ever filing a duplicate. Run this on a cron schedule every two hours.
Start with the Re:amaze List Conversations operation to pull conversations created or updated since the last run. Cover the last two hours with a little overlap so nothing slips through between runs. For each candidate conversation, use List Messages scoped to that single conversation to read the full back and forth, because the opening message often lacks the detail that reveals whether something is actually broken.
Then judge each conversation. A genuine product defect means the product is behaving incorrectly, erroring, or failing to do what it promises. Do not escalate how-to questions, feature requests, billing and refund issues, shipping or delivery complaints, or password and account access problems that are not caused by a defect. If a conversation is genuinely ambiguous, leave it alone rather than filing a speculative bug. Skip conversations that already carry the escalated tag so repeat runs do not reprocess them.
Before creating anything in Linear, search for an existing issue. This step is the point of the workflow, so never skip it. Use the Linear Search Issues operation with the distinctive terms from the report, such as the exact error text, the affected feature, and the page or screen name. Try more than one phrasing, since the customer's wording and the engineer's wording rarely match. Judge whether a result covers the same underlying problem rather than requiring identical words.
If a matching issue already exists, do not create a second one. Use Add Comment to Issue instead, recording that another customer has just hit this, the date, a link to the Re:amaze conversation, what the customer said in their own words, and the running impact such as how many separate reports have now come in.
If the bug is new, use Create Issue. Give it a specific title naming the thing that is broken, for example "Discount codes rejected at checkout for UK customers", never something vague like "Customer reports a bug". In the description include reproduction steps drawn from the customer's own words with the relevant lines quoted directly, what they expected compared with what actually happened, any environment details they mentioned such as browser, device, plan, or order number, and a link back to the Re:amaze conversation. Set priority from how many customers have hit it and how severe it is: Urgent for data loss, broken checkout or payments, or an outage hitting many customers; High for a core feature broken with no workaround, or several separate reports; Normal for a single report with a workaround; Low for cosmetic problems.
Close the loop on the support side. Post the Linear issue link and a one line summary back onto the Re:amaze conversation using Create Message as an internal note, so it is visible to the support team but never sent to the customer. Then use Update Conversation to add an escalated tag to that conversation.
Finally, send a short Slack message with the Slack Bot Send a Message operation to the on-call engineering channel for each escalation. Keep it to a few lines: what is broken, whether this is a new issue or another report of a known one, the priority, and the Linear link. If nothing qualified as a bug in a given run, stay silent rather than posting an empty update.
Example output
What does this prompt do?
- Reads your recent Re:amaze conversations every two hours, following the whole thread rather than just the opening message.
- Separates genuine product bugs from how-to questions, billing queries, and shipping complaints, so engineering only hears about real defects.
- Checks Linear before filing anything. If the problem is already tracked, it adds the new report and the customer impact as a comment, so you get one issue per bug instead of one per complaint.
- Writes new issues with reproduction steps in the customer's own words, links the issue back on the support conversation as an internal note, tags the conversation as escalated, and alerts your on-call channel in Slack.
What do I need to use this?
- A Re:amaze account with access to your support conversations
- A Linear workspace, and the team you want bug reports filed into
- A Slack workspace with a channel where your on-call engineer will see alerts
- A rough sense of how you rank severity, so the agent knows what counts as urgent for your product
How can I customize it?
- Change how often it runs. Every two hours suits a busy inbox, while once or twice a day is plenty for a quieter one.
- Adjust what counts as worth escalating, for example ignoring cosmetic issues or only escalating problems that block a purchase.
- Set your own priority rules, choose which Linear team receives the issues, and pick the Slack channel or person who gets pinged.
- Rename the tag applied to escalated conversations so it matches the labels your support team already uses.
FAQs
Will it open a new ticket every time a different customer reports the same bug?
How does it tell a real bug from an ordinary support question?
Does the customer see any of this?
Do we have to tag conversations as bugs first?
How does it decide the priority of a new issue?
Related templates
When someone leaves, we check which shared passwords they used in their final months and post a ranked rotation list to your security channel.
Every Monday, check every S3 bucket for public exposure, missing encryption and weak backup settings, then get the risks ranked in Slack.
Twice every weekday, the conversations from your social inbox land on the right HubSpot contact timelines, with a Slack recap for sales.
Every morning, find the addresses that hard bounced or filed a spam complaint, update the matching HubSpot contacts, and post a short Slack recap.
Every weekday at 4pm, spot the threads that went quiet, stage a ready-to-send nudge in your mailbox, and get a ranked Slack recap.
When you merge a fix in GitHub, this agent checks the matching dead-letter queue, replays the failed messages, and reports back on the pull request and in Slack.
Stop letting real bugs die in the support inbox.
Set this running and every genuine bug report reaches engineering as one tracked Linear issue, with the customer impact attached.