This implements a structured hypothesis evaluation framework that moves ideas through stages of logical verification, evidence validation, and trust auditing before making a decision. You feed it a problem statement and it orchestrates multiple Claude agents in parallel to generate competing hypotheses, then progressively filters them through L0, L1, and L2 directories based on whether they pass each gate. Everything lands in a .fpf folder with markdown files tracking context, evidence, and a final decision rationale record. It's heavy process for sure, but if you're making architecture decisions or debugging complex system issues where you need to explicitly compare multiple explanations and document why you picked one, the formality actually helps. The parallel agent execution for verification steps is clever.

skills.sh installs
1.2K
GitHub stars
1.8K
last updated
Aug 21, 2026

Install

$ npx -y skills add neolabhq/context-engineering-kit --skill propose-hypotheses --agent claude-code

Installs into .claude/skills of the current project.

Files

Propose Hypotheses Workflow

Execute the First Principles Framework (FPF) cycle: generate competing hypotheses, verify logic, validate evidence, audit trust, and produce a decision.

User Input

Problem Statement: $ARGUMENTS

Workflow Execution

Step 1a: Create Directory Structure (Main Agent)

Create .fpf/ directory structure if it does not exist:

mkdir -p .fpf/{evidence,decisions,sessions,knowledge/{L0,L1,L2,invalid}}
touch .fpf/{evidence,decisions,sessions,knowledge/{L0,L1,L2,invalid}}/.gitkeep

Postcondition: .fpf/ directory scaffold exists.


Step 1b: Initialize Context (FPF Agent)

Launch fpf-agent with sonnet[1m] model:

  • Description: "Initialize FPF context"
  • Prompt:
    Read ${CLAUDE_PLUGIN_ROOT}/tasks/init-context.md and execute.
    
    Problem Statement: $ARGUMENTS
    
    **Write**: Context summary to `.fpf/context.md`**
    

Step 2: Generate Hypotheses (FPF Agent)

Launch fpf-agent with sonnet[1m] model:

  • Description: "Generate L0 hypotheses"
  • Prompt:
    Read ${CLAUDE_PLUGIN_ROOT}/tasks/generate-hypotheses.md and execute.
    

Sponsors

Details

Repository
github.com/neolabhq/context-engineering-kit
Categories
Data Science & ML, Productivity & Planning
First seen
Jun 3, 2026

More from neolabhq/context-engineering-kit

All 63
1alirezarezvani avatarsenior-computer-visionalirezarezvani/claude-skillsComputer vision engineering skill for object detection, image segmentation, and visual AI systems. Covers CNN and Vision Transformer architectures, YOLO/Faster 1.2Kskills.sh installs
2mlflow avatarmlflow-onboardingmlflow/skillsOnboards users to MLflow by determining their use case (GenAI agents/apps or traditional ML/deep learning) and guiding them through relevant quickstart tutorial1.2Kskills.sh installs
3lijigang avatarljg-maplijigang/ljg-skills一个行业 → 一张生态地形图卡(PNG)。以《千脑智能》参考系理论为地基:把行业摊成一张可俯瞰的「生态地形」——价值像河一样流过地貌,再在地形上标出两处——「瓶颈」(流量/产能在此收窄的隘口/水坝)和「价值捕获点」(利润在此沉淀的宝藏堆)。地形让权力结构一眼可见:卡流量的地方常常不是钱沉淀的地方。配三个关键指标的 ba1.2Kskills.sh installs
4qodex-ai avatarlegal-document-analyzerqodex-ai/ai-agent-skillsBuild agents for legal document analysis, contract review, and compliance checking. Handles document parsing, risk identification, and legal research. Use when 1.2Kskills.sh installs
5ruvnet avatarswarm-advancedruvnet/rufloAdvanced swarm orchestration patterns for research, development, testing, and complex distributed workflows1.2Kskills.sh installs
6duneanalytics avatarduneduneanalytics/skillsDune CLI for querying blockchain and on-chain data via DuneSQL, searching decoded contract tables, managing saved queries, managing visualizations, managing das1.2Kskills.sh installs