Fpt Reconstruct

If the user didn't provide a project name, list available projects under first-principals-thinking/projects/ and ask which one to continue.

Read all three artifacts. Confirm:

• Problem statement defines a function, not a form
• Challenge log has summary statistics
• At least one assumption is marked VALIDATED

If the challenge log shows zero validated assumptions, the decomposition needs more work — return the user to /fpt-decompose.

2. Validate First Principles (Six Tests)

For each VALIDATED assumption from the challenge log, apply all six validation tests from references/reconstruction-guide.md § Validation Checklist:

Promote principles that pass all six. Reject or demote those that fail. Document each decision.

Fill out the First Principles Foundation using assets/first-principles-foundation-template.md

Check principle set completeness: physics/mechanics + economics + human/user dimensions all covered?

Save to first-principals-thinking/projects/{project-name}/04-first-principles-foundation.md

3. Reconstruct the Solution

Using only the validated principles, build a new solution:

Phase A — Define the solution space: Map constraints (what principles prohibit) and opportunities (what principles enable). Identify the theoretical ceiling and floor.

Phase B — Generate candidates: Create 2–3 solution candidates. Each must address the functional need from the problem statement, respect every validated principle, and not import structural elements from the old solution unless a principle justifies them.

Use the reconstruction patterns from references/reconstruction-guide.md § Reconstruction Patterns:

• Material Decomposition (cost problems)
• Functional Decomposition (design problems)
• Empirical Re-derivation (performance problems)
• Physical Principle Substitution (mechanism problems)
• Cross-Domain Recombination (innovation problems)

Phase C — Evaluate candidates: Score each against the functional need, principle compliance, feasibility, and testability.

Fill out the Reconstructed Solution using assets/reconstructed-solution-template.md

Every solution element must have a Principle Traceability entry. If an element cannot trace back to a principle, it is either an unexamined assumption (challenge it), a pragmatic concession (acknowledge it), or a creative leap (add to validation plan).

Save to first-principals-thinking/projects/{project-name}/05-reconstructed-solution.md

4. Produce Decision Brief

Synthesize the full cycle into a concise decision brief:

• Problem (one paragraph)
• Key principles discovered (bulleted)
• Recommended solution with rationale
• What this changes from conventional approach
• Risks and validation plan
• Recommended next action

Use assets/decision-brief-template.md

Save to first-principals-thinking/projects/{project-name}/06-decision-brief.md

Return the decision brief inline to the main conversation, plus:

• Next step: "Run /fpt-evaluate {project-name} to score this cycle and log feedback."

Output Format

Three artifacts added to the existing project directory:

1. 04-first-principles-foundation.md
2. 05-reconstructed-solution.md
3. 06-decision-brief.md

Plus the decision brief returned inline with next-step guidance.

Error Handling

• If project directory doesn't exist: list available projects, or instruct user to run /fpt-decompose first
• If decomposition artifacts are missing: list what's missing, instruct user to complete /fpt-decompose
• If zero assumptions passed validation: return user to /fpt-decompose Step 3 for more rigorous challenging
• If user drifts back toward the old solution during reconstruction: flag it — ask which principle justifies that element
• If principle set completeness check fails (dimension missing): pause reconstruction, identify the gap, add it to the foundation

Do Not

• Do not begin without a valid project directory and completed decomposition artifacts
• Do not accept a solution element that cannot trace to a principle — make unexamined imports visible
• Do not generate only one solution candidate — the reconstruction phase benefits from comparison
• Do not skip the decision brief — it is the primary deliverable for stakeholders
• Do not evaluate the cycle's effectiveness — that is /fpt-evaluate's job