Agent-driven triage of a failed BigQuery to Attio audience sync

Turns a raw sync-failure alert into a diagnosed, ticketed problem. When a BigQuery-to-Attio audience sync underdelivers, an agent pulls the rejected rows, looks for the common failure pattern (malformed emails, missing required fields, dedupe collisions), writes a plain-English root cause, and opens a Linear issue with the evidence attached.

Use it when sync failures are frequent enough to drown your data team in manual triage but varied enough that a fixed rule cannot explain them. The agent reads the actual reject reasons instead of just reporting a count mismatch, so the ticket arrives with a hypothesis rather than a shrug.

1. 1A webhook fires when the sync run reports fewer loaded records than extracted.
2. 2The agent queries BigQuery for the run's rejected-record table and reason codes.
3. 3It cross-checks a sample of failing rows against Attio's field requirements.
4. 4Reasoning over the reject reasons, it identifies the dominant failure mode and drafts a root-cause summary.
5. 5It files a Linear ticket with the counts, top reject reasons, and sample failing rows.
6. 6It posts a Slack note linking the ticket so the owner sees it in context.