110+
Azure Cosmos DB best-practice rules
Agent skills for real-world Azure Cosmos DB decisions
Azure Cosmos DB Agent Kit packages practical rules, deterministic scenarios, and evaluation workflows so your agents ship faster, spend fewer RUs, and avoid costly modeling mistakes.
Azure Cosmos DB best-practice rules
Core test scenarios for multi-language iterations
End-to-end tests and automated tests
Critical guidance for partition strategy, data modeling, and query tuning is prioritized so agents make correct architectural choices before code generation drifts.
Scenario contracts and stable test harnesses make regressions visible quickly and reduce noise across iteration comparisons.
Skills follow the Agent Skills format, making the same guidance reusable across Copilot, Claude Code, and other compatible tools.
The cosmosdb-best-practices skill includes tactical guidance for SQL API usage, vector search setup,
full-text search, consistency, and operational diagnostics.
Use the templates in testing-v2/scenarios/ to define SCENARIO.md and API contracts.
Create iteration folders, generate implementation code, and track per-iteration outcomes.
Classify failures, add/adjust rules, rebuild AGENTS docs, and update improvement logs.
npx skills add AzureCosmosDB/cosmosdb-agent-kit
npm install
npm run validate
npm run build
Extend Agent Kit capabilities with the MCP Server and plugins for your favorite AI development tools.
Access Azure Cosmos DB agent skills and best practices via the Model Context Protocol (MCP) standard.
Use Azure Cosmos DB Agent Kit directly in Claude to get best-practice guidance when building and reviewing code.
Integrate Agent Kit intelligence into Cursor for seamless agent-assisted Azure Cosmos DB development workflows.
Available across other AI development platforms and agent frameworks compatible with Agent Skills format.
Share what is working and where you want deeper guidance. Submitting opens a prefilled GitHub issue in this repository.