Luker

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.

canon_scout method — on-demand web search for fanfiction / canon-derived sessions (original-source canon, established fanon, character profiles).

chat_scout method — recent chat scan for unresolved threads, in-flight setups, character states, with signal-vs-noise filtering.

Character voice consistency rules — living-being principle, archetype handling, voice-register matching.

Sub-agent dispatch protocol — task brief shape, parallel pattern, what sub-agents see, briefing discipline.

Anti-meta-narration rules — no config-label leakage, no platform-frame leakage (turn / round / user / system prompt references).

Output length / format discipline — concise reasoning, no idle rounds, parallel dispatch over sequential.

Director main-agent 7-step turn workflow — pre-draft scouting wave, brainstorming, draft, critics, integrate, housekeeping, finalize.

Chinese RP language baseline — bilingual signals are Chinese-first; in-world time frames use Chinese phrasings.

Main-agent draft-writing rules — living-being prose, no meta-narration, three pre-finalize self-check passes.

epistemic_scout method — cross-source per-character knowledge boundary mapping (Knows / Doesn't-know / Omniscience-traps) to prevent POV breaches.

intent_scout method — cross-source extraction of user asks + lorebook authoring directives for this turn.

lorebook_scout method — single-source lorebook scan for setting / worldbuilding / character-canon entries relevant to the scene.

notes_curator method — post-draft housekeeping that closes deployed notes and opens new ones only on genuine new obligations (default = do nothing).

notes_pickup_scout method — scan OPEN notes for trigger-ripe items this turn (planted setups paying off, pending promises being asked about).

plot_brainstormer method — angle-driven structural sketch for the next beat (tension, moves, turning point, foreshadow payoffs).

Preset-scoped test

Character-scoped test

Global test skill A

Global test skill B