generating Architecture Decision Records (ADRs) in MADR or Nygard format from decision context and constraints,
deploying compute workloads — Lambda serverless functions, ECS Fargate containers, EC2 virtual machines,
AWS AI services — Bedrock (Claude, Llama, Titan), SageMaker, Rekognition, Comprehend, Transcribe.
AWS compute skill. Use when: (1) deploying serverless functions with Lambda — event-driven, pay-per-invocation,
setting up a cloud contact centre — instance creation, phone numbers, hours of operation,
AWS managed databases — RDS (PostgreSQL/MySQL), DynamoDB, Aurora, ElastiCache (Redis).
AWS messaging — SQS queues, SNS topics, EventBridge event bus, Kinesis data streams.
AWS networking — VPC design, Route 53, CloudFront CDN, ALB/NLB load balancers, VPC endpoints.
AWS security skill. Use when: (1) designing IAM policies — least-privilege roles, resource-based policies,
AWS storage — S3 object storage, EBS block storage, EFS file system, Glacier archival.
network infrastructure — IOS/IOS-XE/NX-OS switches and routers, VLANs, spanning tree, routing protocols (OSPF, EIGRP, BGP),
Cisco network automation — Netmiko/NAPALM Python, Ansible, DNA Center REST API, NSO, YANG/NETCONF/RESTCONF.
Cisco collaboration — Webex cloud, CUCM on-prem telephony, Unity Connection voicemail, UCCX contact centre.
Cisco switching and routing — IOS/IOS-XE/NX-OS, VLANs, STP, inter-VLAN routing, OSPF, EIGRP, BGP, QoS.
Cisco network security — ASA/Firepower NGFW, Umbrella DNS security, ISE identity/NAC, AnyConnect VPN.
Cisco wireless — Catalyst Wi-Fi (WLC), Meraki cloud-managed, access point deployment, RF design.
setting up a new Claude Code environment for a client or project,
building or deploying Cloudflare Workers or Pages applications,
edge inference — Workers AI for fast, cheap classification and routing,
routing all Claude API calls through AI Gateway for token metering and logging,
building semantic search over skills, documents, or knowledge bases,
running LLM inference at the edge with zero cold starts (Llama, Mistral, Phi),
recording token usage, request counts, or latency metrics per client/skill/agent,
designing schemas and writing migrations for D1 databases,
building real-time collaborative features requiring WebSocket connections,
triggering AI workflows from inbound email (client briefs, intake forms),
connecting Cloudflare Workers to an external Postgres or MySQL database,
storing session tokens, auth state, or user preferences at the edge,
deploying frontend apps, marketing sites, or documentation from a Git repo,
decoupling AI inference from HTTP request handlers to avoid timeout,
storing files, images, documents, or exports uploaded by users,
connecting 2nth Workers to a client's on-premise ERP, CRM, or database without opening firewall ports,
building edge API handlers, middleware, or proxies in TypeScript/JavaScript,
building a platform where YOUR customers deploy their own Workers into your infrastructure,
orchestrating multi-step AI agent pipelines that must survive Worker restarts,
design systems — tokens, theming, dark/light modes, the consistency layer,
building a Discord bot that receives slash commands and responds via Interactions Endpoint,
generating realistic speech from text — TTS with 70+ languages, emotional control, streaming,
automating repositories — create, update, branch, tag via REST API,
authenticating any Google API — gcloud CLI, service accounts, workload identity, OAuth 2.0 with PKCE,
deploying compute — Cloud Run (serverless containers), Cloud Functions (events), GKE (Kubernetes), Compute Engine (VMs),
GCP AI/ML skill. Use when: (1) running Gemini models (gemini-2.5-pro, gemini-2.5-flash, gemini-2.0-flash) via Vertex AI,
GCP compute skill. Use when: (1) deploying serverless containers on Cloud Run — scale-to-zero, pay per request,
GCP data platform skill. Use when: (1) querying or loading data in BigQuery — serverless analytics, per-TB on-demand or capacity,
GCP database skill (stub). Use when: (1) running managed Postgres/MySQL/SQL Server on Cloud SQL,
GCP networking skill (stub). Use when: (1) designing VPC networks — subnets, routes, firewalls, VPC peering,
GCP security skill (stub). Use when: (1) modelling Cloud IAM policies — principals, roles, conditions, policy-simulator,
GCP storage skill (stub). Use when: (1) storing objects in Cloud Storage — buckets, lifecycle, versioning, signed URLs,
reading, sending, or labelling email through the Gmail API,
Workspace Admin SDK skill (stub). Use when: (1) provisioning users and groups via Directory API,
Google Calendar API skill (stub). Use when: (1) creating events via events.insert with Google Meet auto-generation,
Google Drive API skill (stub). Use when: (1) creating, reading, or updating files/folders via Drive API v3,
Gmail API skill. Use when: (1) reading, searching, or listing messages/threads via users.messages.list + get,
Google Sheets API skill (stub). Use when: (1) reading cell ranges or whole sheets via values.get / values.batchGet,
building WhatsApp Business automation — receiving and sending messages via Cloud API,
receiving and processing inbound Instagram DMs in a Cloudflare Worker,
receiving and processing inbound WhatsApp messages in a Cloudflare Worker,
calling Azure OpenAI, Azure AI Agent Service, or Azure AI Foundry from Cloudflare Workers,
calling Azure OpenAI (GPT-4o, o3, embeddings) from Cloudflare Workers via REST,
running open-weight models locally — Llama 3, Mistral, Gemma, Phi, Qwen, DeepSeek with zero cloud cost,
calling Ollama from TypeScript/Python/curl — generate, chat, embed endpoints,
running Ollama in Docker — containerised inference with GPU passthrough,
baking a system prompt permanently into a model — so callers don't need to pass it,
choosing the right open-weight model for a task — size vs quality vs speed trade-offs,
building a Slack bot or app that receives events and replies in channels or DMs,
monitoring mentions and keywords in real-time via the filtered stream in a Cloudflare Worker,
reading or writing CRM data — Leads, Contacts, Accounts, Deals via Zoho CRM API v8,