By wme3
Skills for AI agents to help humans make high-stakes decisions through systematic research and structured analysis
Generate a compelling case FOR a completed decision
Start a new decision process from the Thesis Gate
Generate a compelling case AGAINST a completed decision
Continue a paused decision process
Review a past decision against actual outcomes
Library of cognitive biases for identifying and correcting systematic errors in reasoning during decision-making
Gate 4 - Consolidate Facts vs Assumptions, update Knowns/Unknowns, create assumption inventory
Gate 5 - Pre-mortem, steel-man opposition, surface biases, second-order effects
Generate compelling cases FOR a decision - advocacy grounded in rigorous analysis
Gate 7 - Document final decision, rationale, and commit to artifact
No model invocation
Executes directly as bash, bypassing the AI model
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
A Claude Code plugin for making high-stakes decisions through systematic research and structured analysis.
I discovered Superpowers by Jesse Vincent while looking for ways to improve my AI-assisted coding workflow. After using it, I was impressed by how it enforced discipline— test-driven development (TDD), debugging workflows, code review patterns—through structured skills rather than ad-hoc prompting.
I wanted to understand how it worked, so I forked it and walked through the architecture. The plugin system, the skill definitions, and the session hooks. I was expecting to find optimized prompts, but I found something much deeper - structure.
That exploration sparked an idea: what if I built something similar for decision-making? Not coding decisions, but the messy, high-stakes business decisions that keep founders up at night. Hiring. Strategy. Investment. The decisions where being wrong is expensive and the right answer isn't obvious.
I built Deliberate Decisions as an experiment in applying the same structured, skill-based approach to decision-making that Superpowers applies to coding. It incorporates Kahneman's System 1/System 2 framework, pre-mortem analysis, and reference class forecasting into a 7-gate process that forces slow, deliberate thinking. I was introduced to Kahneman in business school and I've found myself going back to his books time and time again (Thinking Fast & Slow).
I'm sharing it in case others find it useful.
Thanks!
-Matt
Decisions fail in the small assumptions, not the big choices.
This framework surfaces the hidden structure of decisions:
Built on Kahneman's System 1/System 2 framework, Deliberate Decisions enforces slow, deliberate thinking through mandatory gates.
# Clone the repository
git clone https://github.com/yourusername/deliberate-decisions.git
# Load as a Claude Code plugin
claude --plugin-dir /path/to/deliberate-decisions
> /deliberate-decisions:decide
I'm starting a new decision process using Deliberate Decisions.
What decision are you facing?
Every decision passes through seven mandatory gates:
| Gate | Name | Purpose |
|---|---|---|
| 1 | Thesis | Define what we're deciding and why |
| 2 | Landscape | Map alternatives, Decision Points, Knowns/Unknowns |
| 3 | Research | Gather evidence, find reference class |
| 4 | Calibration | Separate Facts from Assumptions |
| 5 | Contrarian | Pre-mortem, steel-man opposition, bias audit |
| 6 | Synthesis | Must-Be-True Conditions, Exit Criteria, recommendation |
| 7 | Decision | Human decides with full picture |
Not all decisions need the same rigor. Choose your depth:
| Weight | Time | Best For |
|---|---|---|
| Light | <10 min | Reversible, domain expertise, lower stakes |
| Medium | 30-40 min | Moderate stakes, some uncertainty |
| Complete | 60-90 min | High stakes, irreversible, significant uncertainty |
Weight affects depth, not gates. All 7 gates always run.
/deliberate-decisions:decide # AI suggests weight after Thesis
/deliberate-decisions:decide --light # Start at Light
/deliberate-decisions:decide --medium # Start at Medium
/deliberate-decisions:decide --complete # Start at Complete
You can upgrade mid-process. The AI will also suggest upgrading if complexity emerges.
| Command | Action |
|---|---|
/deliberate-decisions:decide | Start a new decision |
/deliberate-decisions:advocate | Generate compelling case FOR |
/deliberate-decisions:detract | Generate compelling case AGAINST |
/deliberate-decisions:review-decision | Compare outcomes to predictions |
/deliberate-decisions:resume-decision | Continue a paused decision |
npx claudepluginhub wme3/deliberatedecisions --plugin deliberate-decisionsA decision operating system for high-stakes choices. Simulates disagreement across 5 independent personas, stress-tests assumptions, and converges on what actually holds up.
Truth-seeking sparring partner for Claude Code. Challenges claims, decisions, and documents through structured dialectical analysis.
A structured toolkit of thinking skills for Claude Code — covering logic, ethics, creativity, decision-making, systems thinking, communication, and more. Each skill is an invokable command that applies a specific reasoning framework to any situation.
A deterministic thinking partner that challenges assumptions and applies 150+ mental models to sharpen decisions, solve problems, and think more clearly. Features orientation detection, cognitive operations framework, and structured diagnostic workflows.
Adversarial thinking partner for founders and executives. Stress-tests plans, prepares for board meetings, navigates hard decisions, and forces honest post-mortems.
Structured thinking methods that counteract LLM reasoning biases during problem exploration — first principles, inversion, constraint manipulation, perspective forcing, analogy search, and more, with user-gated parallel subagent exploration for deep dives