Turn tagged Dixa bug reports into clean Linear issues

Every 20 minutes, scan Dixa for customer conversations tagged as bugs and turn each new one into a well-formed engineering issue in Linear, then link that issue back onto the Dixa conversation so support agents can follow the fix.

Let me configure two things up front: which Dixa tag marks a conversation as a bug or defect (for example 'bug', 'defect', or 'regression'), and which Linear team new issues should be filed under. Use those two values throughout the run.

To find candidates, use Dixa's Search conversations with filters to pull recent conversations, and use List conversation tags on each one to confirm it actually carries the watched bug or defect tag. Only continue with conversations that genuinely have that tag.

Before filing anything for a conversation, call List internal notes on it and check whether it already contains a Linear issue link. If a Linear link is already present, skip that conversation so the same bug is never filed twice.

For each new tagged conversation, use List messages to read the full thread and gather the reproduction details and customer impact. Do not invent steps that are not supported by the conversation; if a detail is missing, say so plainly in the issue rather than guessing.

Draft a clear engineering issue: a concise, specific title, and a description that includes steps to reproduce, expected versus actual behavior, and how many customers appear to be affected. File it in Linear with Create Issue under the configured team.

Finally, close the loop: post the new Linear issue back onto the originating Dixa conversation with Add internal note. Keep the note short, with the issue title and its Linear URL, so support agents can follow the fix from inside Dixa.