Quantum Reasoning Skill — documentation

Start here, then pick the depth you need.

First steps

  1. Getting started — what the skill is, how to install it, and the first prompt to try.
  2. Quickstart — a runnable verification script, its verified output, and the provenance of every number it prints.
  3. Skill contract — the normative protocol. If the skill and a doc disagree, SKILL.md wins.

Protocol and implementation

Document What it answers
Skill contract What the model must do, when, and what it must never claim
Measurement methodology The exact formulas, weights and thresholds behind the reference controller
Host compatibility Which host capabilities are required, expressed as capabilities rather than vendors
Branch controller The reference implementation of scoring, dormancy, revival and collapse
Release contract gate What CI asserts about the built distribution, and how the gate proves it is fail-closed
quantum-reasoning CLI The console script that validates the installed skill asset

Examples

Document What it shows
Usage examples Prompt shapes for scientific, mathematical, debugging and planning tasks
Host integration Native skill directory, persistent-instruction wrapper, controller-assisted agent, benchmark A/B setup

Benchmarks

Document What it shows
Benchmark protocol Cases, schemas, evaluator, controls, submission rules
Seed cases Six deterministic smoke-test cases

No baseline-versus-skill model evaluation has been published. The seed cases prove the harness runs; they are not evidence that this skill improves reasoning. Every contributed number must ship the provenance listed in CONTRIBUTING.md.

Project governance

Document Purpose
Contributing How to propose a change and how to submit benchmark results
Code of conduct Expected behaviour in issues, pull requests and discussions
Security policy How to report a vulnerability, and what is in scope
Changelog Release history
Citation metadata How to cite this protocol and software
License MIT

Reporting problems