From tp-fpf
Execute the complete First Principles Framework cycle — generate competing hypotheses, verify logic, validate evidence, audit trust, and produce a decision with full traceability.
How this skill is triggered — by the user, by Claude, or both
Slash command
/tp-fpf:fpf-proposeWhen to use
Use when the user says 'first principles', 'hypothesize', 'propose options', 'FPF', 'evaluate from first principles', 'reason from scratch', or 'generate hypotheses'. IMMEDIATELY when the user asks to 'analyze this problem', 'evaluate alternatives', 'compare solutions', 'make a decision with rationale', or 'think through this carefully'. BEFORE committing to a major technical decision, architectural choice, or complex problem solution.
The summary Claude sees in its skill listing — used to decide when to auto-load this skill
IF analyzing a problem from first principles -> START with context initialization, THEN generate -> verify -> validate -> audit -> decide
IF analyzing a problem from first principles -> START with context initialization, THEN generate -> verify -> validate -> audit -> decide IF user provides problem statement -> Use it directly as the hypothesis focus IF combining with root cause analysis findings -> Use as input evidence IF combining with reflection output -> Use reflection findings to validate hypotheses
The First Principles Framework (FPF) uses the ADI cycle (Abduction-Deduction-Induction): generate competing explanations, verify logical validity, validate against empirical evidence, audit trustworthiness, then decide with full traceability.
.fpf/ directory structure if missing:mkdir -p .fpf/{evidence,decisions,sessions,knowledge/{L0,L1,L2,invalid}}
.fpf/context.mdLaunch sub-agents to generate competing hypotheses (3-5 recommended). Each hypothesis covers a different region of the solution space:
.fpf/knowledge/L0/{id}.mdDisplay hypothesis table, ask user if they want to add their own.
For each L0 hypothesis, launch a sub-agent to:
For each L1 hypothesis, launch a sub-agent to:
For each L2 hypothesis, launch a sub-agent to:
.fpf/evidence/Launch a sub-agent to:
.fpf/decisions/Present results, ask user if they agree with the decision. If not, iterate from step 7.
| Path | Contents |
|---|---|
.fpf/context.md | Problem context and scope |
.fpf/knowledge/L0/*.md | Initial hypotheses |
.fpf/knowledge/L1/*.md | Verified hypotheses |
.fpf/knowledge/L2/*.md | Validated hypotheses |
.fpf/knowledge/invalid/*.md | Rejected hypotheses |
.fpf/evidence/*.md | Evidence and audit files |
.fpf/decisions/*.md | Design Rationale Record |
.fpf/ directory structure exists.fpf/context.md.fpf/decisions/Provides UI/UX resources: 50+ styles, color palettes, font pairings, guidelines, charts for web/mobile across React, Next.js, Vue, Svelte, Tailwind, React Native, Flutter. Aids planning, building, reviewing interfaces.
Fetches up-to-date documentation from Context7 for libraries and frameworks like React, Next.js, Prisma. Use for setup questions, API references, and code examples.
npx claudepluginhub git-fg/taches-principled --plugin tp-fpf