fpf-propose

Decision Router

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

Propose Hypotheses

Core Principle

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.

Process

Step 1: Initialize Context

1. Create .fpf/ directory structure if missing:

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

2. Launch a sub-agent to initialize context: read the problem statement, analyze scope, write context to .fpf/context.md

Step 2: Generate Hypotheses (Parallel)

Launch sub-agents to generate competing hypotheses (3-5 recommended). Each hypothesis covers a different region of the solution space:

• High-probability candidates (the obvious explanations)
• Low-probability alternatives (creative but plausible) Each written to .fpf/knowledge/L0/{id}.md

Step 3: Present Summary + User Input

Display hypothesis table, ask user if they want to add their own.

Step 4: Verify Logic (Parallel per hypothesis)

For each L0 hypothesis, launch a sub-agent to:

1. Verify internal logical consistency
2. Check for hidden assumptions
3. Move to L1 (verified) or invalid

Step 5: Validate Evidence (Parallel per hypothesis)

For each L1 hypothesis, launch a sub-agent to:

1. Search for supporting or refuting evidence in the codebase
2. Cross-reference with existing knowledge
3. Move to L2 (validated) or invalid

Step 6: Audit Trust (Parallel per hypothesis)

For each L2 hypothesis, launch a sub-agent to:

1. Calculate R_eff (evidence reliability score)
2. Identify weakest link (WLNK)
3. Write audit report to .fpf/evidence/

Step 7: Decide

Launch a sub-agent to:

1. Review all L2 hypotheses and audit reports
2. Create Design Rationale Record (DRR) in .fpf/decisions/
3. Present recommendation with rationale

Step 8: Present Final Summary

Present results, ask user if they agree with the decision. If not, iterate from step 7.

Artifacts Created

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

Completion Checklist

• .fpf/ directory structure exists
• Context recorded in .fpf/context.md
• Hypotheses generated, verified, validated, and audited
• DRR created in .fpf/decisions/
• Final summary presented to user