| name | final-method-explainer |
| description | Build the authoritative final method explanation for a submission-ready subquestion from the method card, human decision ledger, code plan, final results, and robustness evidence. |
Purpose
Explain the selected method completely without re-authoring why the human chose it.
Preconditions
rigor_profile is submission or writer handoff is explicitly requested.
- Human method choice and result verdicts are recorded in
qx_decisions.jsonl.
- Approved code ran and final result/robustness evidence exists.
Sources
Use:
qx_method_card.md
qx_decisions.jsonl
planning/model_assumptions.md
planning/symbol_table.md
qx_code_plan.md
- final run summary and result analysis
- robustness summary/report
Read legacy candidate and iteration logs only for migration.
Workflow
- Resolve the final method and baseline from the latest non-stale human decisions.
- Transcribe the human's selection rationale faithfully and cite its
decision_id.
- Explain:
- goal and scope;
- assumptions, including human-confirmed necessity labels;
- symbols and units;
- mathematical formulation;
- inputs, outputs, objective/criteria, and constraints;
- solution procedure;
- baseline and why it is valid;
- fallback trigger and whether it fired;
- validation, robustness, limitations, and applicable range.
- Ensure formulas match code and symbol table.
- Save
methods/Qx/qx_final_method_explanation.md.
Rules
- Do not infer the chosen method from best metrics.
- Do not invent the why-this-method narrative.
- Do not create a new pending decision artifact.
- Do not restate a long iteration diary; include only material eliminated alternatives and evidence.
- Do not include unsupported numerical claims.
Verification
- Final method and rationale trace to decision IDs.
- Assumptions, symbols, formulas, units, and code agree.
- Baseline is usable rather than merely diagnostic.
- Risks, fallback behavior, and limitations are explicit.
- The explanation is self-contained enough for the writer.