What problem does it solve? Application Insights log exports are rich but noisy, making it hard to pinpoint why a specific request is slow. This Skill parses the CSV export, ranks performance candidates (slow requests, slow SQL, N+1 patterns, app-time gaps), traces the chosen issue to the exact executing code in your .NET repository, and guides a scoped fix through to a linked Azure DevOps work item and PR. ## Core Features & Use Cases - Structured CSV Parsing: A bundled Python parser handles multiline quoted SQL fields and outputs JSON with per-request durations, normalized SQL signatures, duplicate query detection, and app-time gap calculations. - Guided Six-Phase Workflow: Validates the repo matches the telemetry namespace, presents a ranked candidate table for the user to pick from, traces the endpoint to the offending LINQ/EF Core code, proposes a scoped diff, runs a regression sweep, and ships via Azure DevOps. - Scope Discipline: Enforces one issue at a time, cross-checks git history before calling fixes incomplete, and keeps work item titles in plain language for product owners and testers. - Use Case: You export an App Insights CSV showing a 35-second request. The Skill identifies a 32-second app-time gap caused by materializing an unfiltered result set, locates the exact LINQ query in your controller's repository, and proposes a filtered Where clause fix. ## Quick Start Analyze the attached Application Insights CSV export and help me find and fix the slowest request in this repository.
