Dakota Smith | Fullstack Solutions Architect

Software Engineer

I build things that build things. Agentic systems, AI tooling, and enterprise digital solutions.

Read My Articles

Featured Post

Latest from the blog

Cover image for Preparing for Mythos: Enterprise CMS Security Checklist

securitycmsweb-developmentdotnet

Preparing for Mythos: Enterprise CMS Security Checklist

Protect your enterprise CMS stack against the vulnerabilities Mythos exposed with this platform-specific checklist for Sitecore, Umbraco, and Optimizely.

April 15, 202611 min read

Learn

A structured reference for agentic engineering

16Patterns

6Chapters

9Deep-Dives

4Sub-pages Each

Patterns

Named, repeatable techniques for AI-assisted engineering. Organized into 6 chapters from foundation to advanced orchestration.

CH.1Foundation4 patterns

CH.2Context4 patterns

CH.3Task4 patterns

CH.4Steering4 patterns

CH.5Verification0 patterns

CH.6Recovery0 patterns

Browse patterns

Toolkit

Expert's guide to Claude Code internals. Mental models, production playbooks, composition patterns, and real-world pitfalls.

CLAUDE.mdProject instruction architecture, layered configs, context priming strategies

HooksExecution lifecycle automation, validation pipelines, agent guardrails

SkillsReusable prompt engineering, skill composition, invocation patterns

Agents & SubagentsDelegation patterns, isolation strategies, worktree workflows

Agent TeamsMulti-agent orchestration, role-based pipelines, coordination architectures

MCP ServersExternal tool integration, custom server patterns, protocol mastery

CommandsCustom slash commands, workflow automation, command composition

Settings & ConfigPermission architectures, model selection strategy, profile management

Memory SystemPersistent context strategies, memory taxonomy, cross-session intelligence

Explore toolkit

Sample patterns

1.1Convention FileBeginnerMake implicit project rules explicit in a file the agent always reads, so it follows your standards without being told each time.

2.1Context PrimingIntermediateFront-load the most relevant files and decisions before making a request, so the agent starts from understanding rather than guessing.

3.1Vertical SliceIntermediateBreak features into thin end-to-end slices rather than horizontal layers, giving the agent complete context for each piece.

Browse Patterns Explore Toolkit View Language Map

Builds

Plugin Architectplugin Plugin Opsplugin Plugin GTMplugin HandoffKitproduct CMS Analyzers Marketplaceplugin APL - Autonomous Phased Looperagent ORCagent Conduitinfrastructure STUDIOagent