Skip to content

Instantly share code, notes, and snippets.

View skew202's full-sized avatar
:octocat:
Good code speaks to machines; great code speaks to humans

steven skew202

:octocat:
Good code speaks to machines; great code speaks to humans
  • ɯoɔ.ןıɐɯƃ@Ƨ0Ƨʍǝʞs
View GitHub Profile
@skew202
skew202 / developer-environment.md
Created January 3, 2026 04:13
Developer Environment & Toolchain. Hardware context (Lenovo Legion, NVIDIA GPU, CUDA), CLI tools (ollama, hf, gh, kubectl, terraform, hcloud, doctl), OSS-exclusive philosophy, monorepo with git submodules for each artifact.

Developer Environment & Toolchain

Author: skew202 Updated: 2026-01-03

The Philosophy

"My agents need to know my environment. Hardware, tools, and constraints all feed into optimal context."

@skew202
skew202 / session-journal-template.md
Created January 3, 2026 04:04
Session Journal Template. Empty and filled templates for AI session logging. Tracks metadata, context, work done, decisions, problems, agent interactions, research notes, next session context, metrics.

Session Journal Template

Author: skew202 Updated: 2026-01-03

Quick Copy-Paste Template




  


  




    


          
    

      

        
        

          

              @skew202
          

          

            
                skew202
                / project-structure.md
            
            

              Created
              January 3, 2026 04:04
            

              
                Project Structure & Session Management. Monorepo with research/, archive/ (session journals), .ai/ (Claude/Cursor/Antigravity configs), hooks, MCP, LSP. Session start/end rituals, context continuity.
              
          

        

      

        

  

      

    
Project Structure & Session Management


Author: skew202
Updated: 2026-01-03




The Philosophy




"Every project is a research endeavor. Every session builds on the last. Every AI agent needs proper context."





  


  




    


          
    

      

        
        

          

              @skew202
          

          

            
                skew202
                / engineering-excellence-checklist.md
            
            

              Created
              January 3, 2026 03:57
            

              
                Engineering Excellence Checklist. Pre-merge verification: C4 docs, Semver, Conventional Commits, testing pyramid (unit>80%), OWASP Top 10, K8s, ArgoCD, CI/CD, observability. Language-agnostic standards.
              
          

        

      

        

  

      

    
Engineering Excellence Checklist: Standards Verification


Author: skew202
Updated: 2026-01-03




Purpose


This checklist ensures every implementation meets industry-leading standards while incorporating bleeding-edge research insights.



  


  




    


          
    

      

        
        

          

              @skew202
          

          

            
                skew202
                / standards-synthesis.md
            
            

              Created
              January 3, 2026 03:57
            

              
                Standards Synthesis: Established Excellence + Bleeding Edge Research. Opus synthesizes GoF/Martin/Fowler standards with latest cs.AI/cs.MA research. C4, Semver, Conventional Commits, 12 Factor, OWASP, K8s, ArgoCD, GitOps, testing pyramid.
              
          

        

      

        

  

      

    
Standards Synthesis: Established Excellence + Bleeding Edge Research


Author: skew202
Updated: 2026-01-03




The Philosophy




"True excellence combines timeless principles with cutting-edge insights. Not just one or the other."





  


  




    


          
    

      

        
        

          

              @skew202
          

          

            
                skew202
                / meta-research-phase.md
            
            

              Created
              January 3, 2026 03:31
            

              
                The Meta-Research Phase: Before Any Code. Pre-B research workflow: Grok (SOTA discovery) → Opus (first principles) → Multi-agent critique → Perplexity (deep research) → Project guide. Replace 'just code' with 'first understand'.
              
          

        

      

        

  

      

    
The Meta-Research Phase: Before Any Code


Author: skew202
Updated: 2026-01-03




The Philosophy




"Implementation is downstream of understanding. First understand, then build."





  


  




    


          
    

      

        
        

          

              @skew202
          

          

            
                skew202
                / multi-agent-patterns.md
            
            

              Created
              January 3, 2026 03:30
            

              
                Multi-Agent System Patterns. Specialist architecture, parallel voting, hierarchical, debate/red-team. Orchestration patterns, state machines, error recovery, communication protocols. LangGraph/CrewAI/AutoGen comparison.
              
          

        

      

        

  

      

    
Multi-Agent System Patterns


Author: skew202
Updated: 2026-01-03
Inspired by: LangGraph, CrewAI, AutoGen patterns




The Multi-Agent Mindset



  


  




    


          
    

      

        
        

          

              @skew202
          

          

            
                skew202
                / spec-template.md
            
            

              Created
              January 3, 2026 03:30
            

              
                Software Specification Template for AI-Native Engineering. Complete 16-section spec template: Problem, Requirements, Architecture (Mermaid), Data Model, API, UI, Edge Cases, Tests, Security, Migration, Monitoring.
              
          

        

      

        

  

      

    
Software Specification Template


Author: skew202
Updated: 2026-01-03
Purpose: Spec-driven development for AI-native engineering




Why Specs Before Code?



  


  




    


          
    

      

        
        

          

              @skew202
          

          

            
                skew202
                / prompt-patterns.md
            
            

              Created
              January 3, 2026 03:30
            

              
                Prompt Patterns for Engineering Agents. ReAct, Chain of Thought, Tree of Thoughts, Few-Shot, Reflexion, Generated Knowledge, Self-Consistency. Domain-specific patterns for frontend/backend/database.
              
          

        

      

        

  

      

    
Prompt Patterns for Engineering Agents


Author: skew202
Updated: 2026-01-03
Inspired by: promptingguide.ai, ReAct Paper




Patterns vs. Prompts



  


  




    


          
    

      

        
        

          

              @skew202
          

          

            
                skew202
                / agent-evaluation.md
            
            

              Created
              January 3, 2026 03:29
            

              
                AI Agent Evaluation & Self-Verification Framework. 5-layer verification: Agent self-check, automated CI/CD, LLM evaluation (DeepEval/RAGAS), cross-agent review, human checklist. Security, performance, metrics.
              
          

        

      

        

  

      

    
AI Agent Evaluation & Self-Verification Framework


Author: skew202
Updated: 2026-01-03
Inspired by: "Specs, Tests, and Self-Verification" (AI Engineering Podcast)




The Discipline Problem