Primitives

agent-chain-configure

Orchestrate builder, reviewer, and tuner agent handoffs for solution plays — define roles, model routing, token budgets, and quality gates

api-rate-limit-configure

Configure API rate limiting, APIM throttling, and retry behavior for AI endpoints — absorb bursts, isolate noisy tenants, and control 429 retries

app-insights-configure

Configure Application Insights, OpenTelemetry, and KQL dashboards for AI workloads — trace latency, token usage, failures, and cost signals

audit-log-implement

Implement immutable audit logging, PII redaction, and Sentinel monitoring for AI systems — preserve evidence, prove access history, and detect suspicious activity

azure-openai-setup

Set up Azure OpenAI deployments, RBAC, and monitoring — harden inference endpoints and control latency, quota, and cost

backup-restore-ai

Back up AI data, vector indexes, and configuration state — recover prompts, conversations, and search assets after failures

bicep-module-create

Create reusable Bicep modules, parameter files, and registry packages — standardize Azure infrastructure and reduce deployment drift

canary-deploy-ai

Implement canary deployment, rollout gates, and rollback automation for AI changes — catch latency or quality regressions before full release

circuit-breaker-add

Add a circuit breaker, fallback routing, and health checks for Azure OpenAI calls — contain 429 and 503 cascades and preserve service

compliance-audit-run

Run AI compliance audits, collect Azure evidence, and track remediation — assess SOC 2, HIPAA, ISO 27001, and EU AI Act controls

embedding-model-select

Select embedding models, dimensions, and hybrid search settings for vector workloads - balance recall, storage cost, and migration risk

fai-agent-chain-configure

Configure builder→reviewer→tuner agent chain with handoff rules, model assignment, and evaluation gates.

fai-agent-governance

Review agent safety controls, budget limits, and human escalation paths - approve only agents that stay bounded, observable, and policy-compliant

fai-agentic-eval

Run evaluator-optimizer loops for agents and prompts - score outputs, revise weak instructions, and stop only when quality thresholds hold

fai-api-docs-generator

Generate API reference docs, examples, and error catalogs from OpenAPI or code annotations - keep endpoint docs accurate, testable, and publishable

fai-api-endpoint-generator

Generate typed REST API endpoints with Zod/Pydantic input validation, RFC 7807 error responses, OpenAPI annotation, and Managed Identity auth — eliminating schema drift and IDOR vulnerabilities.

fai-api-rate-limit-configure

Configure API rate limiting with Azure APIM policies, throttling tiers, and retry headers.

fai-app-insights-configure

Configure Application Insights with custom AI metrics, anomaly alerts, and operational dashboards.

fai-architecture-blueprint

Generate solution architecture blueprints with Mermaid service topology, component responsibility tables, data flow annotations, and WAF pillar mapping — bridging system design to implementation.

fai-architecture-decision-record

Generate Architecture Decision Records capturing context, forces, decision rationale, rejected alternatives, and consequences — preventing knowledge rot when teams diverge from original design intent.

fai-aspire-orchestration

Wire .NET Aspire AppHost for multi-container orchestration, automatic service discovery, distributed OpenTelemetry export, and local Azure emulator integration — eliminating hardcoded connection strings and manual container start order.

fai-aspnet-minimal-api

Scaffold ASP.NET Core Minimal API with typed route groups, FluentValidation, Managed Identity credentials, OpenAPI v3 via Scalar, Serilog structured logging, and health check endpoints — cutting AI service setup time from hours to minutes.

fai-audit-log-implement

Implement immutable audit logging with Azure Blob Storage, Sentinel integration, and compliance evidence.

fai-az-cost-optimize

Analyze Azure AI workload spend with Cost Management API queries, flag overprovisioned PTU, identify semantic cache gaps, right-size model SKUs, and produce actionable FinOps recommendations with ROI estimates.