Can I add my own hypotheses to the process?

Yes, the workflow includes an interactive step where you can input your own hypotheses for the agents to verify, validate, and audit alongside AI-generated ones.

How does this skill use sub-agents?

It launches parallel 'fpf-agents' to verify logic and validate evidence for each hypothesis independently, ensuring a thorough and unbiased analysis.

What is the First Principles Framework (FPF)?

FPF is a methodology that breaks down complex problems into basic truths to build up logical, evidence-based solutions from the ground up, avoiding assumptions.

What is a Design Rationale Record (DRR)?

A DRR is a standardized document generated at the end of the cycle that summarizes the final decision, the rationale behind it, and the supporting evidence.

Where are the FPF results stored?

All data, including hypotheses, evidence, and final decisions, are stored in a local .fpf/ directory within your repository for transparency and version control.

First Principles Hypothesis Proposer

Name: First Principles Hypothesis Proposer
Author: NeoLabHQ

byNeoLabHQ

•

687

•

Productivity & Workflow

Execute the complete First Principles Framework cycle to transform problem statements into evidence-backed decisions using autonomous agents.

The First Principles Hypothesis Proposer automates the First Principles Framework (FPF), a rigorous reasoning methodology designed to solve complex engineering and architectural problems. It orchestrates a multi-step workflow where specialized agents generate competing hypotheses, verify logical consistency, validate empirical evidence, and audit the reliability of sources. By creating a structured .fpf/ knowledge base, the skill ensures that decisions are fully documented with a Design Rationale Record (DRR), providing a transparent, audit-ready decision-making process for developers and teams.

Key Features

01Comprehensive trust auditing with R_eff scoring and weakest link identification

02687 GitHub stars

03Automated multi-agent hypothesis generation and verification

04Automated generation of Design Rationale Records (DRR) for project documentation

05Structured directory-based knowledge management using the .fpf/ format

06Parallel logic verification and evidence validation using high-context models

Use Cases

01Evaluating competing architectural patterns for a new software system

02Systematic debugging of complex or intermittent system failures

03Evidence-based technology stack selection and justification

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add neolabhq/context-engineering-kit propose-hypotheses

For use in Claude.ai and ChatGPT

Download Skill