Skip to content

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

52 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Andie v6.4

What is Andie? Andie turns your AI into a room of sharp advisors who argue, pressure-test your choice, and hand you a decision worth shipping.

Claude vs Andie: With Claude alone, you get one voice agreeing with you. With Andie, you get a room that disagrees — until the right choices survive.

Why Andie? You review code before shipping — Andie does the same for decisions, so high-stakes calls get pressure-tested before you commit.

Life without Andie: lone-wolf choices, hand-wavy plans, regret. With Andie: plan first, debate, gate, ship — get it right, not just fast.

Runs on Claude Code, Claude.ai, ChatGPT, Gemini, Perplexity, Copilot, Grok, Manus, and any LLM. MIT. Built by Giggso.


How to Install Andie

  • Claude Code: npx andie-skill — installs the skill. Restart Claude and /andie to begin.
  • Claude.ai / Claude Desktop: download dist/andie.skill and upload it as a skill.
  • ChatGPT, Gemini, Perplexity, Copilot, Grok, Manus, or any LLM: copy the matching file from dist/ into the platform's system prompt — see dist/README.md for the per-platform paste target.
  • Plain link install: share https://github.com/giggsoinc/andie — every install path lives in this one repo.

How to Work with Andie

Say one word to pick a mode, or just describe your problem and Andie picks for you.

📘 Deep — teacher at the whiteboard

  • Use when you want to understand something, not decide on it.
  • Triggers: "explain", "teach me", "how does X work", "deep on Y".
  • Andie loads a Functional, Technical, and Data expert and unpacks it layered.
  • Every round ends with a Feynman recap — plain-language insight you can repeat back.
  • Output: a clear mental model and the next question worth asking.

🎭 Drama — writers' room debate

  • Use when you face a tradeoff or contested choice — architecture, vendor, strategy.
  • Triggers: "should I", "X vs Y", "which approach", "help me decide".
  • Named experts argue by name; a Devil's Advocate seat is always staffed.
  • You hold the casting vote at every gate; Andie never decides for you.
  • Output: a scored recommendation with the dissent recorded so you ship eyes-open.

🚨 War — incident commander

  • Use when production is degrading or down right now and time is the constraint.
  • Triggers: "outage", "down", "broken in prod", "urgent", "triage".
  • Skips pre-flight, runs T+minutes log, gives imperative actions only.
  • One commander voice, blast-radius first, restore before root-cause.
  • Output: stabilization plan now, post-incident handoff to Kaizen later.

🔄 Kaizen — improvement detective

  • Use for recurring failures, process drift, or "why does this keep happening".
  • Triggers: "keeps breaking", "root cause", "5 whys", "fishbone", "improve".
  • Picks the right method: Kaizen Cycle, Ishikawa, 5 Whys, DMAIC, Pareto, or A3.
  • Evidence chains and numbered cycles — no hand-waving, no jumping to fixes.
  • Output: a confirmed root cause and one fix at a time, never a shotgun.

License

MIT. Use it, fork it, ship it.

About

Multi-modal sharp thinker — FeynTech + Drama modes for Claude, Perplexity, ChatGPT, Manus

Topics

Resources

Contributing

Stars

5 stars

Watchers

0 watching

Forks

Releases

Packages

Contributors

Languages