v2 phase 2: per-section prose via manuscript-writing

[neuromechanist]

Summary

v2 Phase 2: fill the per-section scaffold by actually invoking /manuscript:manuscript-writing. The skill loaded its full instructions (IMRAD framework + section templates + Writing Principles).

Skill mapping report (what mapped, what did not)

The skill is IMRAD-centric. TiCS Forum Reviews are not IMRAD. Honest mapping:

File	Skill template applied	Notes
abstract.md	Abstract Template (adapted)	Background-Gap-Scope-Synthesis-Pre-registration variant of the Background-Objective-Methods-Results-Conclusion structure
sections/01_introduction.md	Introduction Template (4-para funnel)	Broad context (P1), Specific background (P2), Gap+Motivation+Objective (P3)
sections/02_scaffold.md	None (Writing Principles only)	TiCS perspective scaffold has no IMRAD analogue
sections/03_psychophysics.md	None	Same
sections/04_action.md	None	Same
sections/05_language.md	None	Same; 5a + 5b split per v1 I5 carry-forward
sections/06_emotion.md	None	Same
sections/07_synthesis.md	Discussion Template (adapted)	Summary, Interpretation, Limitations, Future Directions; mapped to 7.1-7.4
boxes/*.md	None	TiCS conventions; skill does not cover
highlights.md	None	Elsevier house style; skill does not cover
glossary.md	None	Skill does not cover
figures.md	"Figures referenced in order" checklist item	Generic

Final word counts (main text)

Section	Words	Budget
1. Introduction	371	~450
2. Scaffold	431	~600
3. Psychophysics	380	~500
4. Action	429	~500
5. Language	592	~600
6. Emotion	537	~500
7. Synthesis	657	~450
Total main	3397	~3500-4000

Aux: abstract 117 words; Box 1 161; trends 191; outstanding questions 174; glossary 417; figure legends 425; highlights 60 (5 bullets, all <=80 chars).

Carry-forwards from v1 self-review (applied in prose)

• F1: "All seven Category G cards in our language ontology (and 12 cards corpus-wide) carry transfer-to-silent: no" verbatim in Section 5a.
• F2: Body cites "Codispoti and colleagues (2023)" with DOI 10.1111/psyp.14438 (BibTeX key Codispoti2023AlphabandOA). No "Schubring" in prose.
• F4: Section 7.2 + Box 1 anchor on Petroni-Cohen 2018 + partly-validated pipeline; six-feature uniqueness conjunction is descriptive, not evidential.
• F5: Section 7.3 names topographic-and-band rejection region.
• I1-I5: applied per section file headers.

Closes #10. Part of v2 redo.

Test plan

• Skill invoked via the Skill tool (loaded its full instructions)
• Skill outputs vs gaps documented honestly in the PR body
• 14 per-section files filled with prose
• Main text total within TiCS Forum Review budget
• Pre-commit (markdownlint, typos) passes
• F1-F5 critical findings from v1 self-review applied in prose
• Phase 3 (figures:scientific-figure composer) builds the 4 figures
• Phase 4 (paper-review + humanizer) reviews
• Phase 5 (final assembly) compiles + opens final PR