What problem does it solve? Debugging a misbehaving service usually means digging through observability dashboards by hand. This Skill lets you query HyperDX logs and traces directly from your AI session using Lucene syntax, against either the HyperDX cloud API or a local ClickHouse instance running in Docker. ## Core Features & Use Cases - Dual-mode querying: Search the HyperDX cloud REST API with a Personal API key, or query a local HyperDX container's ClickHouse directly via docker exec. - Lucene query support: Use field filters like level:error, service:name, and TraceId lookups, plus free-text search, with multiple --query terms automatically OR'd together. - Trace inspection: Switch to the traces table in local mode to inspect spans, durations, and status codes for distributed tracing. - Use Case: A user notices elevated 500 errors on their API. They ask the assistant to check the logs, and the Skill runs hdx_query.sh with level:error filtered to their service over the last 30 minutes, then summarizes the repeated error patterns. ## Quick Start Ask the assistant to check HyperDX for recent errors on your service, for example: show me any level:error logs from the api service in the last 15 minutes.