Play 29
MCP Gateway
Centralized MCP server management with rate limiting and governance.
A governance layer for MCP servers. Azure API Management acts as a gateway proxy — authenticating requests, rate limiting per client, load balancing across MCP server instances, and collecting usage analytics. Tool discovery aggregates capabilities from all registered MCP servers into a unified catalog. GPT-4o-mini handles intelligent tool routing (matching user intent to the right MCP tool). Full observability via Azure Monitor dashboards.
Architecture Pattern
MCP gateway: APIM proxy, tool discovery, intelligent routing, usage analytics
Azure Services
DevKit (.github Agentic OS)
- agent.md — root orchestrator with builder→reviewer→tuner handoffs
- 3 agents — MCP Gateway Builder (gpt-4o), Reviewer (gpt-4o-mini), Tuner (gpt-4o-mini)
- 3 skills — deploy (101 lines), evaluate (103 lines), tune (100 lines)
- 4 prompts — /deploy, /test, /review, /evaluate with agent routing
- .vscode/mcp.json — FrootAI MCP with backend API key input + envFile
TuneKit (AI Config)
- config/openai.json — gpt-4o-mini for cost-efficient routing
- config/gateway.json — rate limits, auth policies, tool registry
- config/guardrails.json — tool access controls, request validation
- evaluation/eval.py — Routing accuracy >90%, p99 latency <200ms
Tuning Parameters
Machine evidence
FrootAI evidence lifecycle
This is an internal evidence maturity label, not third-party certification, accreditation, legal compliance, or a production guarantee. Missing or expired evidence demotes automatically; catalog claims cannot promote a play.
This play currently has design evidence only. A runnable scenario, endpoint evaluation, and build receipts are the next contiguous gates.
Repo Intelligence
v1A no-clone, revision-pinned map for agents and humans. Observed evidence is separated from inferred flow so the output stays useful without pretending to be a full call graph.