Testing the Azure Cosmos DB Agent Kit: what it found reviewing a Claude-built app, in RUs saved
adnan_hashmi · x · 2026-10-23
Benjamin Bork ran a controlled experiment: he had Claude Opus 5.5 (high effort) build an "Audio Blog" app from a single prompt — paste an article URL, get a natural-sounding MP3 — then used the Azure Cosmos DB Agent Kit to review the resulting code.
- Fixed stack: ASP.NET Core minimal API + background worker, React/Vite/MUI frontend, Cosmos DB for NoSQL, Azure OpenAI for text rewriting, Azure AI Speech neural TTS, Blob Storage for MP3s, deployed on Container Apps / Static Web Apps with Bicep IaC and managed identity.
- To make the test fair he disabled the Microsoft Learn and Azure MCP servers, isolating the Agent Kit's contribution.
- The post details what the Agent Kit flagged and how many RUs the fixes actually saved; demo source is open on GitHub (benjaminbork/azure-cosmos-db-agent-kit-demo).
More from coding & agent
- PHP hit by two unauthenticated RCE flaws; ProjectDiscovery launches free internet-wide alert service — cyb3rops · 2026-10-23
- Modal open-sources Spindle, a Tinker-compatible training server that charges GPU time instead of tokens — josh_wills · 2026-10-23
- Dev ditches endless chat threads for a visual dashboard to manage her AI agents — Salmaaboukarr · 2026-10-23
- Warp cuts coding agent verification bottleneck with per-agent computer-use and clip reviews — vikvang1 · 2026-10-23
- Computer Use Is Solved; Specifying Goals Is the New Bottleneck for Agents — alishbaimran_ · 2026-10-23
- Decisions API hits GA as agents automate cloud video editing via feedback loops — dkundel · 2026-10-23