folio:plan

folio:plan — Mode-Specific Planning

Generate planning artifacts based on the confirmed mode. Read route.json to determine the effective mode (user_override takes precedence over selected_mode).

Entry Conditions

• route.json exists with a confirmed mode.
• Canonical inputs available under inputs/.
• Human Checkpoint 1 and routing checkpoint approved (check logs/checkpoints.md).

If missing, direct user to run folio:prep first.

Protocol by Mode

Reference ../../references/modes/*.md for detailed mode protocols where available.

---

White Paper (W2 + W3)

W2: Perspective Discovery

1. Scan inputs/raw_materials/, idea.md, and available external literature.
2. Write planning/perspectives.json: array of {perspective_id, viewpoint, key_arguments, source_materials, relevance}. At least 3 distinct perspectives (or user waiver).
3. Write planning/question_tree.json: structured questions grouped by planned section.

W3: Argument Structure

1. Formulate thesis and strongest counterthesis in planning/argument_graph.md.
2. Map claims to perspectives; plan ledger support for narrative sections.
3. Write planning/argument_graph.md — argument flow: sections, claims, dependencies.
4. Write planning/claim_ledger.json with extended types: fact, interpretation, opinion, forecast. Conform to ../../templates/manifests/claim_ledger.schema.json.

Gate C — Claim ledger:

python ../../scripts/build_claim_ledger.py workspace/

Block on failure; fix until pass.

Exit: perspectives.json, question_tree.json, argument_graph.md, claim_ledger.json exist. Gate C passes.

---

Research Paper (D2)

D2: Global Planning

1. Read canonical inputs thoroughly.
2. Generate planning/outline.json — title, sections array (id, name, purpose, key_points, estimated_length, depends_on_figures, depends_on_citations), narrative_arc.
3. Generate planning/claim_ledger.json — claims with id, statement, type, evidence_source, support_level, depends_on_figures, depends_on_citations, notes.
4. Generate planning/figure_plan.json — per figure: id, description, type, source, source_path, supports_claims, caption_draft.
5. Generate planning/literature_plan.json — search_queries, known_references, coverage_areas with min ref counts.
6. Generate planning/results_inventory.json — results entries: id, description, value, source_file, supports_claims, verified.

Gate B — Plan completeness:

• Outline includes standard venue sections.
• Every claim has support_level; unsupported claims explicitly flagged.
• Figure plan covers outline-referenced figures.
• Literature plan covers citation needs.

Human Checkpoint 2: Present outline, narrative arc, claim ledger summary (highlight weak/unsupported), figure and literature plans. Wait for explicit approval.

Exit: All 5 planning artifacts exist. Gate B passes. Checkpoint 2 approved.

---

Hybrid (H2)

Follow W2 + W3 for the conceptual front-half:

• planning/perspectives.json
• planning/question_tree.json
• planning/argument_graph.md
• planning/claim_ledger.json (extended types)

Gate C:

python ../../scripts/build_claim_ledger.py workspace/

Exit: Same as white paper planning artifacts. Gate C passes.

Next Stage

• White paper: proceed to folio:draft (white paper skips folio:support).
• Research paper: proceed to folio:support.
• Hybrid: proceed to folio:support.