agentleFS
wangsalinwangsalin/ai-acting-system/skills/01-心理动机/character-memory/SKILL.md

character-memory

Skill2 starsChanged 2 months agoMITView on GitHub

角色长线记忆与场景状态转移。跨镜头、跨场次维护心理状态演进与情绪余波。

The file

Files
SKILL.md

Character Memory Skill (角色记忆与连续性状态)

skills/character-memory.md 负责维护角色在多场次、多镜头、跨剧集之间的动态记忆与状态继承,杜绝长篇 AI 剧集中角色“每场戏重新做人”的脱节现象。


1. 核心职能

在真实影视表演中,演员进入新的一场戏时,必然携带着上一场戏的生理残留与心理余波。

本模块跟踪记录:

  1. Emotional Residue (情绪余波与衰减):刚经历激烈争吵、惊吓或欢愉后,生理肾上腺素不会瞬间归零,面部张力、呼吸深浅、视线警惕度会有残留。
  2. Physical Continuity (身体与物理状态):体能消耗、是否有外伤、受冻/出汗、衣物整洁度、发型凌乱度、左右手持物。
  3. Information & Secrets (信息差与信任动态):本场戏角色新获悉了什么事实?开始怀疑谁?隐藏了什么新秘密?
  4. Unfinished Actions (未完成动作意志):上一场戏被打断的动作(例如正欲倒水、正欲点烟、正欲开门)。

2. 状态记忆数据格式

{
  "character_id": "char_01_linwan",
  "current_scene": 14,
  "previous_scene": 13,
  "time_elapsed": "30分钟前(争吵刚结束)",
  "physical_state": {
    "posture": "身体僵硬,两肩轻微耸起",
    "breathing": "略微短促,未完全恢复平缓",
    "held_items": ["右手握着手机(屏幕熄灭)"],
    "costume_condition": "大衣纽扣扣错一颗,头发有一缕散在脸颊旁"
  },
  "emotional_residue": {
    "dominant_residue": "被欺骗后的麻木与应激性冷淡",
    "decay_rate": "缓慢,需至少经历一次有效安抚或新的外部重大刺激才会化解",
    "guard_level": "极高(对男主的所有主动接触表现出防御性避让)"
  },
  "cognitive_state": {
    "just_learned": "男主瞒着自己买了去往北方的单程车票",
    "unspoken_suspicion": "男主可能早在三个月前就已经做出放弃关系的决定",
    "strategic_posture": "不再歇斯底里,转为冷眼观察与测试对方谎言"
  }
}

3. 跨场景状态转移规则

规则 1:无重大事件触发,情绪绝不发生跳变

  • 错误案例:第 12 场结尾大哭分手崩溃,第 13 场开头无过渡直接若无其事地和同事开怀大笑。
  • 正确修正:第 13 场开头即便是在面对同事,角色的笑容也是晚半秒挤出来的“社交面具”,同事转身后眼神立刻跌落回空洞,嘴角肌肉垮塌。

规则 2:追踪“道具与物理锚点”

角色上一镜头拿在手里的手机、钥匙、水杯,进入下一镜头必须有交代(仍在手中、随手放在玄关桌上、塞进风衣口袋),避免出现 AI 生成常见的“道具凭空消失或瞬间换手”。

规则 3:信息沉淀改变台词节奏

当角色获悉某个秘密后,即使台词不变,其反应时间、视线聚焦方式也会发生剧变:从原本的“即时回应”变为“先凝视对方眼角半秒,再给出简短回答”。

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.