What problem does it solve? Convex deployments accumulate hidden performance and cost problems — read limit hits, OCC write contention, and repeated function failures — that are hard to spot without digging through logs. This Skill reads the deployment's own 72-hour insights events and traces each one back to the exact line of code causing it, turning runtime symptoms into evidence-backed, fixable findings. ## Core Features & Use Cases - Live insights analysis: Reads typed 72h health events (documentsRead/bytesRead limits and thresholds, OCC retries and permanent failures) via the official Convex MCP tools. - Code-level root-causing: Opens the flagged function's source to identify unindexed queries, missing pagination, .collect() scans, or read-modify-write hotspots, citing file and line. - Structured findings with fixes: Emits severity-ranked findings (perf/cost/correctness) with concrete repairs such as adding an index with .withIndex, .paginate, sharded counters, or aggregate components. - Use Case: Your Convex app's messages:list query is slowing down and costs are rising. Run the advisor to discover it read 4.2MB from the messages table 31 times yesterday due to a missing index, and get the exact fix applied on confirmation. ## Quick Start Ask the assistant to check my Convex deployment's health insights and root-cause any performance or contention issues in my functions.
