harshilmathur

harshilmathur / autodecision

Public

A decision operating system for high-stakes choices — business, strategy, career. Simulates disagreement, stress-tests assumptions, and converges on what actually holds up. Claude Code skill inspired by Karpathy's autoresearch + LLM council.

23
0
100% credibility
Found Apr 20, 2026 at 23 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Python
AI Summary

A tool for Claude AI that generates structured decision briefs for high-stakes choices by simulating expert personas, stress-testing assumptions, and producing robust recommendations.

How It Works

1
🔍 Discover the decision helper

You hear about a smart tool that helps make big life or work choices by thinking like a team of experts.

2
📱 Add it to your AI chat

In your favorite AI conversation app, you easily add this helper with a simple search and one-click install.

3
💭 Ask about your tough choice

You type something like 'Should we change our prices?' and the helper springs into action, breaking it down step by step.

4
🤝 Watch it debate and test ideas

It gathers real facts, imagines different viewpoints like optimists and critics, spots risks, and refines until everything agrees.

5
Choose quick check or full deep dive
🚀
Quick mode

Get a speedy summary of pros, cons, and advice right away.

🔬
Full mode

Dive deep with multiple rounds of testing and a complete report.

📄 Receive your decision guide

You get a clear, beautiful report with recommendations, risks, and why they make sense, ready to share or act on.

Sign up to see the full architecture

4 more

Sign Up Free

Star Growth

See how this repo grew from 23 to 23 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is autodecision?

Autodecision is a decision operating system built as a Claude plugin for high-stakes choices in business, strategy, and career moves. It decomposes your question, grounds it with web searches, simulates disagreement via five parallel personas—like Growth Optimist and Risk Pessimist—then stress-tests assumptions through critiques, adversary scenarios, and sensitivity analysis to produce a structured Decision Brief. Implemented in Python with Claude Code or Cowork integration, it stores runs locally as a github decision log for tracking outcomes.

Why is it gaining traction?

Unlike basic github decision tree python tools or simple LLM prompts, autodecision enforces a rigorous pipeline inspired by autoresearch and LLM councils, forcing second-order effects, black swans, and convergence checks that solo reasoning misses. Users get shareable briefs with sourced probabilities, fragility thresholds, and export options to PDF or Notion, plus modes like /autodecision:quick for 2-minute runs or :challenge for red-teaming proposals. The hook is interactive iteration until stability, surfacing synthesized options no single persona spots.

Who should use this?

Execs acting as chief operating decision maker (CODM) under IFRS 8 or US GAAP for operating decision making, like pricing cuts or M&A. Founders evaluating build-vs-buy, market expansion, or fundraising terms. Career switchers weighing relocation or role shifts where first-order thinking fails—skip for low-stakes or high-conviction calls.

Verdict

Early alpha with 23 stars and 1.0% credibility score signals low maturity, but polished docs, real examples (law firm AI pilots, buy-vs-rent), and battle-tested validation make it worth trying for serious github decision making. Install via Claude marketplace; pair with your own decision tree from scratch for custom tweaks.

(198 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.