菜单
Chinese RP language baseline — bilingual signals are Chinese-first; in-world time frames use Chinese phrasings.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Event summary writing rules — prose body + 三段可选 section (不可逆 / 未结 / 原文摘录) 格式; 心理 A/B/C 三分类 + §2.1 颗粒度准则 (default 一句过场 + 三类例外允许展开) + §2.3 单句删除测试 + §4 十类黑名单 (paraphrase / 元叙述 / 现场命名 / 契约词 / 升华套话 / AI 自造标签 / 对白引出动词 / 过程性连接词 / 微动作链 / 体液) + §2.6 NPC baseline 路由到 character_sheet。本规范同时为 leaf event 写作规范, 以及 rollup compress 必须遵守的*写作纪律*。
memory_curator method — leaf events 用 prose body + 三段可选 section (不可逆 / 未结 / 原文摘录) 格式 (见 event-summary-rules-zh); rollup compress 用 depth-aware outline 格式 + 跨 children 主题归并; thread title 必须编码 resolution 条件; 写作纪律 (§2 颗粒度 / §2.2 心理 A/B/C / §4 十类黑名单) leaf 与 rollup 共享。
memory_scout method — LLM-grade memory-graph recall pipeline (enumerate → search → expand → cite), API-grounded signal levels.
Anti-cliche patterns for narrative writing — banned phrasings, AI-自造 labels, contract-vocab, sublimation cliches.
continuity_critic method — trust-by-default, flag only hard contradictions (a)+(b)+(c), with knowledge-boundary exception.
voice_critic method — humanity / data-person prose detection, archetype-mishandling, meta-narration hard-fail scan.
| name | director-zh-style-baseline |
| description | Chinese RP language baseline — bilingual signals are Chinese-first; in-world time frames use Chinese phrasings. |
| metadata | {"author":"Luker Team","version":"1.0.0"} |
This skill is the Chinese-RP style baseline for the default director RP profile. The default profile targets Chinese-language RP as the primary use case, and many of the pattern lists in the sub-agent prompts (especially voice_critic) are bilingual but list the Chinese form first. This skill consolidates the Chinese-specific markers so all agents share the baseline.
The baseline is extracted verbatim from director-default-prompt.js and director-defaults.js (voice_critic + EVENT_SUMMARY_RULES_BODY).
When the original prompts list a pattern to avoid (cold observation verb, data vocabulary, meta-narration leak, etc.), the Chinese form is typically the primary form and the English is the fallback. Examples:
观察 / 分析 / 推测 / 记录 / 评估 / 追踪 / 监测 / 扫描 / 检测 / 实验 / observe / analyze / measure / record / monitor / track / scan / log / experiment
心率 / 体温上升 / 充血程度 / 多巴胺 / 肾上腺素 / 皮质醇 / 效率 / 任何百分比 / heart rate up / dopamine / cortisol / efficiency / any % readout
"[角色]注意到 X" / "[角色]记录到 Y" / "第 N 次发生 Z" / "[character] noted that X" / "[character] observed Y dispassionately" / "for the Nth time"
"[角色]像在观察珍稀动物一样" / "用陈述事实的语气" / "冷静地指出" / "with clinical detachment"
Author-side config labels in Chinese: 世界书 / 角色卡 / 记忆图 / 笔记 — all of these are author-substrate names that must NOT appear in-prose. Common leakage shapes:
Turn / round / reply structure used as a time anchor (the most common Chinese-RP failure mode):
Conversation-as-artifact references:
Platform / RP framing:
User-as-referent:
Interface references:
When the narrator looks back at past events, use IN-WORLD time frames:
Never use platform frames:
Sanity test: would the in-world character have a concept for this time anchor? A noble at her dressing table has 昨夜 / 今早 / 三天前 but not 上一回合. A swordsman has 上次相遇 / 雨停那一刻 but not 上次回复. If the time anchor only makes sense relative to the AI-player conversation, it is platform-frame leakage.
The event-summary writing rules (see event-summary-rules-zh) forbid 「」 / "" quoted dialogue inside summary text. In narrative prose this is allowed, but be aware that all 8 ASCII / CJK quote variants are tracked as dialogue markers when the voice_critic / memory_curator scan looks for paraphrase residue or quoted dialogue leaking into structured fields.
When evaluating a candidate phrasing, ask in Chinese: "这句话读起来像一个真实人物在经历这一刻吗,还是像一台保安摄像头在记录?" (Does this line read like a living being having this moment, or like a security camera recording it?) — the data-person failure mode is the most common cliche in Chinese-RP, and this is the single sharpest filter.
