aetherbound-guild/docs/runtime/p9_1/visual_slice/HANDOFF.md

P9.1 Formation-To-Diagnosis Visual Handoff

Scope

This handoff covers one representative chain only:

formation review -> approach/contact -> damage -> heal response
-> ranged finish -> first causal diagnosis

It is a design and layer reference, not integrated runtime source and not Owner visual acceptance. Open index.html?frame=<id> at exactly 844x390.

Representative Truth

• All four accepted Recruit identity anchors are included in assets/actors/; this representative pair renders only Farstring and Kindhand and does not imply that either is a fixed protagonist.
• Starting pair: Farstring + Kindhand (FS-KH).
• Automatic order: Farstring is order 0, exposed-right lead; Kindhand is order 1, protected-left second member.
• Battle 1 starts both at 100 HP; exposed Farstring takes -42 to 58; Kindhand restores +12 to 70.
• The final remote response is a representative ordered event showing the enemy at 34/100; it is not presented as a new combat formula.
• Result 1 names the system order, both profession actions, current HP, the once-only +20 Coin receipt and the next adaptation comparison.

Frames

ID	Contract proof	First read
formation-review	P91-E04	Automatic order and both responsibilities
approach-contact	P91-E05 contact	Actual target geometry before impact
damage-feedback	P91-E05 damage	One -42, attached to Farstring, 100 -> 58
heal-response	profession response	Kindhand-to-Farstring connector and one +12
ranged-finish	ranged response	Farstring-to-enemy direction and one -30
first-diagnosis	Result 1	Earliest divergence, response and cause-first copy

Editable Layers

The HTML keeps layers semantic and independently replaceable:

1. .layer-environment: accepted hand-painted battlefield and scene wash.
2. .layer-actors: stable Recruit/enemy art and action poses.
3. .layer-connectors: party cord, target line, projectile and heal path.
4. .layer-feedback: signed number and semantic VFX.
5. .layer-ui: status, objective, context rail and controls.
6. .layer-captions: ordered event ribbon or cause-first result text.

Runtime implementation should bind those layers to immutable party snapshots and ordered events. It must not reproduce branch logic from these static frames.

Accessibility And Variants

• Every mechanic has text and geometry in addition to color.
• The current target remains visible before impact.
• Captions use 来源 -> 动作 -> 目标 -> 结果.
• Touch controls are at least 48x48; primary actions are 56 high.
• Reduced motion can use the same six stable poses and connectors with no camera travel, parallax, bobbing or continuous motion.
• Reduced flashes can remove any transient emphasis without removing contours, signed numbers or before/after values.
• Mute changes no geometry, timing or caption.

Integration Notes

• Preserve accepted actor sources and their stable IDs; do not redraw names or profession labels into raster art.
• Do not copy the static result values into UI logic. Bind resolved result clauses from P91_RESULT.
• Do not expose manual order, front/rear toggles or combat commands.
• The result context rail is deliberately dense at 100% and is not a 130% acceptance frame. Runtime 130% reflow follows the frozen spec.

Residual Gates

• Owner visual judgment remains open.
• Runtime comprehension, motion feel, sound, devices, packaging and release remain outside this visual-slice handoff.