Hybrid Mode¶

Hybrid mode combines executive narrative framing with evidence-heavy sections. It runs both the white paper conceptual pipeline and the research paper support artifact pipeline, merges them into a single draft, and applies a dual review covering both conceptual quality and evidence quality.

When to use¶

Choose hybrid mode when your manuscript needs to:

• Open with a narrative or strategic frame, then back it up with empirical results
• Target an audience that includes both executives and technical reviewers
• Combine a positioning argument with reproducible evidence
• Produce both an executive summary and a submittable research artifact

route_mode.py recommends hybrid when materials include a significant mix of strategic documents and experimental/data artifacts.

Stage H2: Conceptual Framing¶

Follows the white paper W2 and W3 protocol. Produces:

• planning/perspectives.json — Perspective map with source materials and relevance
• planning/question_tree.json — Structured questions grouped by planned section
• planning/argument_graph.md — Argument flow: sections → claims → dependencies
• planning/claim_ledger.json — Extended claim types (fact, interpretation, opinion, forecast)

Gate C — Claim ledger validation:

cd plugins/folio && python scripts/build_claim_ledger.py workspace/

Stage H3: Evidence Support¶

Follows the research paper D3 protocol (substages 3A, 3B, 3C). Merges evidence mapping into the existing claim ledger — uses a single claim_ledger.json with a section or track field to distinguish conceptual from evidence claims where the schema allows.

Produces:

• Updated planning/claim_ledger.json (merged conceptual + evidence tracks)
• citations/citation_pool.json and citations/refs.bib
• figures/captions.json, figures/generated/, figures/supplied/
• tables/generated/ as needed

Gates C, D, and E apply as in research paper mode.

Stage H4: Merged Outline and Draft¶

Folio produces a single planning/outline.json covering both conceptual and evidence sections, with section track metadata (conceptual | evidence) where helpful.

Drafting format — user choice at Checkpoint 2:

• Option A (split): Conceptual sections in drafts/paper.md; evidence sections in drafts/paper.tex; cross-references documented in drafts/README_hybrid.md
• Option B (unified): One format only — full Markdown or full LaTeX — user decides

check_artifacts.py runs on any LaTeX portions before review.

Human Checkpoint 2: Folio presents the merged outline and, optionally, partial draft sections. You approve the structure and format choice before full drafting completes.

Stage H5: Dual Review¶

Both the conceptual and evidence tracks must pass quality gates. There is no averaging — a failure in either track blocks the stage.

IP scan (required):

cd plugins/folio && python scripts/scan_redlines.py workspace/

IP Gate: Violations block completion.

Hostile review: Conceptual parts → reviews/hostile_review.md (logical gaps, unsupported assertions, competitive risk, clarity)

Evidence review: Data-heavy parts → reviews/review_round_N.md and reviews/scorecard.json with the standard six dimensions.

Gate G — Non-regression: Same rule as other modes; max 3 rounds; best-scoring draft retained.

Stage H6: Combined Export¶

Human Checkpoint 3: Folio presents the final scorecard, deferred issues, sensitive content, and the split-vs-unified artifact plan. Explicit approval required.

Artifacts assembled:

• paper.md and/or paper.tex
• exports/executive_summary.md
• exports/bd_talking_points.md
• citations/refs.bib
• figures/ contents
• reviews/ip_safety_report.md
• reviews/hostile_review.md
• Scorecards

Compile LaTeX portions:

cd plugins/folio && bash scripts/compile_package.sh workspace/

Final bundle assembled by package_exports.py under final/ and/or exports/.

Artifacts produced¶

See Artifacts for the complete cross-mode artifact list.