Skip to content

T8mars/minimax-h3-prompt-skill-T8 / skills/identity-clue-reconnection

identity-clue-reconnection

Verified

Create original short videos using the 身份线索回连|局部累积到主体确认 Creative DNA for MiniMax H3 or Seedance 2.0. Use when the user wants this causal, camera, motion, rhythm, or payoff structure with new subjects and surfaces.

FileKindSize
SKILL.mdskill md2.3 KB
agents/openai.yamlconfig354 B
references/summary.mdmarkdown887 B
references/template.mdmarkdown4.3 KB

references/template.md

View on GitHub
# 身份线索回连|局部累积到主体确认 Template

## Mechanism

Build one stable identity through partial but mutually consistent visual clues; after they accumulate, return to a recognizable head/face marker already glimpsed earlier. The source never supplies a final full-body proof, and the face is not fully concealed at the start. An original transfer may strengthen the closing proof without claiming the source did so.

## Invariants

### inv-01

- Rule: Recurring visual design markers across different close crops must remain attributable to one stable character.
- Purpose: A later identity echo can only resolve the clue sequence if the earlier clues have one common owner.
- Failure if removed: If details come from different unmarked subjects, the ending cannot support an identity connection and becomes an arbitrary montage.

### inv-02

- Rule: The sequence must alternate several body regions while repeating at least one design marker; not every crop has to introduce new information.
- Purpose: The middle sustains familiarity and variation before the final identity echo.
- Failure if removed: If the montage stays on only one undifferentiated crop, the later eye/head echo no longer gathers dispersed details into one recognizable design.

### inv-03

- Rule: The ending returns to a previously glimpsed identity marker and preserves recurring design cues, without requiring a source full-body view.
- Purpose: Provides a recognizable identity echo after the detail ladder rather than an unrelated final insert.
- Failure if removed: If the ending drops the earlier identity marker and recurring cues, the fragment sequence no longer resolves to a recognizable stable character.


## Variable Slots

| Slot | Constraint |
|---|---|
| subject_identity | One original figure with a stable design; a human version must be clearly adult, but the mechanism also works with a non-human figure. |
| clue_objects | Three or more original details attributable to one subject; a transfer may reconnect them more explicitly than the source. |
| setting | One new environment with a stable direction or landmark; avoid the source's blank-field anime design. |
| camera_grammar | Close detail views and a final recognizable identity echo; an optional wider whole-subject proof is an original strengthening, not a source claim. |
| decisive_action | An optional original role-confirming action can strengthen the new instance without being attributed to the source. |
| sound_family | Original, action-linked sound design; no claim that the source music or effects were heard. |

## Required Human-Facing Anchors

- 同一主体
- 局部线索相互一致
- 结尾回连已见标记

## Anti-Copy Boundary

- Do not reproduce the source anime character's face, red eyes, black strappy costume, gold hair ornaments or body proportions.
- Do not copy the original exact crop order, cut timings, palette or bare-white visual field.
- Do not use creator names, franchise identifiers, source caption wording or creator prompt text.

## Transfer Tests

1. An adult offshore lighthouse maintainer is identified through salt-stained gloves, a wrist compass and safety line before activating a beacon.
   - Preserved: inv-01, inv-02, inv-03
   - Changed: subject_identity, clue_objects, setting, camera_grammar, decisive_action, sound_family
2. A stop-motion brass courier automaton is introduced through a stamped wheel, folded paper wing and blue-glass eye before its head returns in a miniature post-office set.
   - Preserved: inv-01, inv-02, inv-03
   - Changed: subject_identity, clue_objects, setting, camera_grammar, decisive_action, sound_family

## Model Skeletons

### MiniMax H3

    integrated_multimodal_description: [new subject, environment, ordered actions, camera, light and final state preserving every anchor]
    overall_soundscape: [diegetic ambience, action cues and causal payoff]
    non_diegetic_music: [original music arc or N/A]

### Seedance 2.0

    任务:[原创实例意图]
    总时长约[4–15]秒。
    主体定义:[稳定人物、物体与空间关系]
    事件或镜头:[按因果顺序实现全部锚点]
    整体风格与画质:[原创媒介、光影、色彩与声音]
    约束:[连续性、物理逻辑、反复制与结尾状态]

Adapt shot count to the user's concept. Preserve the causal mechanism, not a fixed storyboard.
slack-gif-creatoranthropics/skillsKnowledge and utilities for creating animated GIFs optimized for Slack. Provides constraints, validation tools, and animation concepts. Use when users request animated GIFs for Slack like "make me a GIF of X doing Y for Slack."Stars: 198kUpdated Claude CodeCodexVerifiedInstall
ponytailDietrichGebert/ponytailLazy senior dev mode for any coding task: the smallest change that fully solves it, and a reply a busy human understands in one read.Stars: 196kUpdated Claude CodeCodexVerifiedInstall
ponytail-debtDietrichGebert/ponytailHarvest every shortcut: comment into one debt ledger, so deferrals get tracked instead of forgotten. One-shot report.Stars: 196kUpdated Claude CodeCodexVerifiedInstall
ponytail-debtDietrichGebert/ponytailList every `shortcut:` comment (and older `ponytail:` ones) as a debt ledger. One-shot report, changes nothing. Use for "ponytail debt", "what did ponytail defer", "list the shortcuts", /ponytail-debt.Stars: 196kUpdated Claude CodeCodexVerifiedInstall