First Principles Framework (FPF)

What it does

FPF (First Principles Framework) is an organised, auditable architecture for complex reasoning. It helps teams and agents decompose problems into bounded contexts, assign roles, evaluate confidence, generate systematic alternatives, and compose disciplined proposals that survive expert scrutiny. The framework provides patterns for provenance, trust/assurance scoring (F-G-R), reasoning cycles, and composition operators for combining partial solutions.

When to use it

Use FPF when confronting high-stakes or cross-domain problems that need explicit boundaries, traceable assumptions, or multi-stakeholder convergence — e.g., architecture decisions, policy design, cross-discipline research, agent orchestration under budget/trust constraints, or auditing claims for reliability. Avoid for trivial task planning or generic brainstorming.

What's included

• Scripts: Yes — repository includes scripts and a structured sections directory for modular reading
• References: No dedicated references folder, but extensive sectional docs and example starter prompts are present
• Instructions: Detailed navigation guidance, router for starting points, stepwise application advice, and a starter prompt for engineers/managers

Compatible agents

Best suited for agents exposed to long-form spec content (Claude Code, Copilot-style assistants, Codex-like tools) and for humans using it as a structured prompt/guide. It can be used by agent orchestrators that understand multi-step reasoning and tool gating.