Artificial Intelligence

AI

AI tools, platforms, and patterns I actively use to build and ship software.

Claude Code

Daily driver for writing, shipping, and maintaining production code

  • Primary development tool for writing, refactoring, and debugging code
  • Custom slash commands and hooks for project-specific workflows
  • MCP server integration for extending Claude with domain-specific tools

Codex

Delegating autonomous tasks and parallel refactoring in the cloud

  • Autonomous task execution in sandboxed cloud environments
  • Parallel task delegation for multi-file changes and refactoring

OpenClaw

Model-agnostic coding with custom tools and local models

  • Model-agnostic: works with OpenAI, Anthropic, and local models
  • Extensible with custom tools and MCP servers

Mastra AI

Building AI agents and workflows with TypeScript

  • Building AI agents with tool use, multi-step workflows, and human-in-the-loop approval
  • Integrating agents into Next.js and Node.js applications with model routing across providers

Ollama

Running local models for private inference and testing

  • Running open-source models locally for development and testing
  • Private inference for sensitive codebases and internal tools

RAG

Grounding LLM responses with real data from docs and codebases

  • Embedding-based search over documentation and codebases
  • Context injection to reduce hallucination and improve accuracy

MCP

Building custom servers in Go to give LLMs access to internal tools and data

  • Building custom MCP servers in Go for domain-specific tooling
  • Exposing internal APIs, databases, and services to LLM agents
  • Integrating MCP with Claude Code, OpenClaw, and other clients

Skills

Building reusable prompt workflows to automate development tasks

  • Custom slash commands that encode project conventions and patterns
  • Automating repetitive development tasks through structured prompts

Agents

Designing multi-step autonomous workflows with tool use and orchestration

  • Multi-step task execution with tool use and error recovery
  • Orchestrating sub-agents for parallel workloads
  • Building agent workflows that integrate MCP, RAG, and custom tools