@aswless_854771076/ai_short_studio_cli 0.1.156 → 0.1.158

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (175) hide show
  1. package/README.md +1 -1
  2. package/dist/commands/canvas/document/get.js +3 -3
  3. package/dist/commands/canvas/document/save.js +8 -2
  4. package/dist/commands/canvas/documents.js +1 -1
  5. package/dist/commands/canvas/workflow/status.js +1 -1
  6. package/dist/domain/canvas-schema-key.d.ts +2 -0
  7. package/dist/domain/canvas-schema-key.js +21 -0
  8. package/dist/domain/vvicat-api.js +1 -1
  9. package/package.json +1 -1
  10. package/skills/using-vvicat-ai-short-studio-cli/SKILL.md +24 -24
  11. package/skills/using-vvicat-ai-short-studio-cli/references/agent-recovery.md +1 -1
  12. package/skills/using-vvicat-ai-short-studio-cli/references/assets-and-storyboard.md +11 -0
  13. package/skills/using-vvicat-ai-short-studio-cli/references/commands/canvas.md +25 -5
  14. package/skills/using-vvicat-ai-short-studio-cli/references/humanizer/SOURCE.md +9 -0
  15. package/skills/using-vvicat-ai-short-studio-cli/references/humanizer/humanizer.md +408 -5
  16. package/skills/using-vvicat-ai-short-studio-cli/references/project-and-canvas.md +1 -1
  17. package/skills/using-vvicat-ai-short-studio-cli/references/short-drama/SKILL.md +17 -31
  18. package/skills/using-vvicat-ai-short-studio-cli/references/short-drama/assets/generate-character-images.md +2 -2
  19. package/skills/using-vvicat-ai-short-studio-cli/references/short-drama/assets/generate-character-profiles.md +2 -2
  20. package/skills/using-vvicat-ai-short-studio-cli/references/short-drama/assets/generate-drama-art-style.md +3 -3
  21. package/skills/using-vvicat-ai-short-studio-cli/references/short-drama/assets/generate-prop-assets.md +2 -2
  22. package/skills/using-vvicat-ai-short-studio-cli/references/short-drama/assets/generate-scene-assets.md +2 -2
  23. package/skills/using-vvicat-ai-short-studio-cli/references/short-drama/assets/manage-drama-art-styles.md +3 -3
  24. package/skills/using-vvicat-ai-short-studio-cli/references/short-drama/assets/manage-drama-assets.md +4 -4
  25. package/skills/using-vvicat-ai-short-studio-cli/references/short-drama/assets/modules/ideate-drama-from-market/prompts/market_inspiration.en.txt +1 -1
  26. package/skills/using-vvicat-ai-short-studio-cli/references/short-drama/assets/modules/ideate-drama-from-market/prompts/market_inspiration.zh.txt +1 -1
  27. package/skills/using-vvicat-ai-short-studio-cli/references/short-drama/assets/modules/plan-drama-production/prompts/production_plan.en.txt +1 -1
  28. package/skills/using-vvicat-ai-short-studio-cli/references/short-drama/assets/plan-drama-assets.md +2 -2
  29. package/skills/using-vvicat-ai-short-studio-cli/references/short-drama/codex-contracts.md +3 -3
  30. package/skills/using-vvicat-ai-short-studio-cli/references/short-drama/creation/analyze-drama-source.md +2 -2
  31. package/skills/using-vvicat-ai-short-studio-cli/references/short-drama/creation/define-drama-brief.md +1 -2
  32. package/skills/using-vvicat-ai-short-studio-cli/references/short-drama/creation/design-drama-bible.md +1 -1
  33. package/skills/using-vvicat-ai-short-studio-cli/references/short-drama/creation/humanizer.md +4 -4
  34. package/skills/using-vvicat-ai-short-studio-cli/references/short-drama/creation/outline-drama-series.md +1 -1
  35. package/skills/using-vvicat-ai-short-studio-cli/references/short-drama/creation/short-drama.md +1 -1
  36. package/skills/using-vvicat-ai-short-studio-cli/references/short-drama/creation/write-drama-director-book.md +1 -2
  37. package/skills/using-vvicat-ai-short-studio-cli/references/short-drama/creation/write-drama-episode.md +1 -2
  38. package/skills/using-vvicat-ai-short-studio-cli/references/short-drama/editing/edit-deliver-drama.md +2 -2
  39. package/skills/using-vvicat-ai-short-studio-cli/references/short-drama/editing/edit-drama-timeline.md +4 -5
  40. package/skills/using-vvicat-ai-short-studio-cli/references/short-drama/editing/remotion-best-practices.md +4 -5
  41. package/skills/using-vvicat-ai-short-studio-cli/references/short-drama/editing-workflow.md +8 -8
  42. package/skills/using-vvicat-ai-short-studio-cli/references/short-drama/feature-completeness.md +25 -25
  43. package/skills/using-vvicat-ai-short-studio-cli/references/short-drama/generation/configure-generation-providers.md +5 -11
  44. package/skills/using-vvicat-ai-short-studio-cli/references/short-drama/generation/design-drama-audio.md +4 -2
  45. package/skills/using-vvicat-ai-short-studio-cli/references/short-drama/generation/drama-generation-service.md +6 -6
  46. package/skills/using-vvicat-ai-short-studio-cli/references/short-drama/generation/generate-drama-videos.md +4 -4
  47. package/skills/using-vvicat-ai-short-studio-cli/references/short-drama/generation/generate-storyboard-images.md +3 -3
  48. package/skills/using-vvicat-ai-short-studio-cli/references/short-drama/generation/monitor-drama-tasks.md +9 -7
  49. package/skills/using-vvicat-ai-short-studio-cli/references/short-drama/generation/transform-drama-media.md +2 -2
  50. package/skills/using-vvicat-ai-short-studio-cli/references/short-drama/generation/write-drama-video-prompts.md +9 -8
  51. package/skills/using-vvicat-ai-short-studio-cli/references/short-drama/market/analyze-drama-market.md +8 -8
  52. package/skills/using-vvicat-ai-short-studio-cli/references/short-drama/market/ideate-drama-from-market/market-inspiration.schema.json +1 -1
  53. package/skills/using-vvicat-ai-short-studio-cli/references/short-drama/market/ideate-drama-from-market.md +1 -1
  54. package/skills/using-vvicat-ai-short-studio-cli/references/short-drama/market/scripts/README.md +59 -0
  55. package/skills/using-vvicat-ai-short-studio-cli/references/short-drama/market/scripts/analyze.mjs +517 -0
  56. package/skills/using-vvicat-ai-short-studio-cli/references/short-drama/market/scripts/analyze.test.mjs +354 -0
  57. package/skills/using-vvicat-ai-short-studio-cli/references/short-drama/market/scripts/juchacha-client.mjs +188 -0
  58. package/skills/using-vvicat-ai-short-studio-cli/references/short-drama/market/scripts/juchacha-client.test.mjs +218 -0
  59. package/skills/using-vvicat-ai-short-studio-cli/references/short-drama/market/scripts/normalize.mjs +458 -0
  60. package/skills/using-vvicat-ai-short-studio-cli/references/short-drama/market/scripts/normalize.test.mjs +236 -0
  61. package/skills/using-vvicat-ai-short-studio-cli/references/short-drama/market/scripts/report.mjs +69 -0
  62. package/skills/using-vvicat-ai-short-studio-cli/references/short-drama/market/scripts/store.mjs +341 -0
  63. package/skills/using-vvicat-ai-short-studio-cli/references/short-drama/market/scripts/store.test.mjs +302 -0
  64. package/skills/using-vvicat-ai-short-studio-cli/references/short-drama/module-map.json +43 -133
  65. package/skills/using-vvicat-ai-short-studio-cli/references/short-drama/music-catalog-providers.md +3 -3
  66. package/skills/using-vvicat-ai-short-studio-cli/references/short-drama/operations/manage-drama-projects.md +5 -5
  67. package/skills/using-vvicat-ai-short-studio-cli/references/short-drama/operations/orchestrate-short-drama.md +7 -8
  68. package/skills/using-vvicat-ai-short-studio-cli/references/short-drama/operations/recover-drama-pipeline.md +1 -1
  69. package/skills/using-vvicat-ai-short-studio-cli/references/short-drama/operations/short-drama-skill-index.md +7 -7
  70. package/skills/using-vvicat-ai-short-studio-cli/references/short-drama/operations/use-short-drama-studio.md +4 -6
  71. package/skills/using-vvicat-ai-short-studio-cli/references/short-drama/pipeline.md +40 -65
  72. package/skills/using-vvicat-ai-short-studio-cli/references/short-drama/project-spec-v1.md +91 -131
  73. package/skills/using-vvicat-ai-short-studio-cli/references/{hypit/playbooks/craft → short-drama/prompt}/h3-prompt-writing.md +4 -4
  74. package/skills/using-vvicat-ai-short-studio-cli/references/short-drama/prompt/prompt-craft.md +186 -0
  75. package/skills/using-vvicat-ai-short-studio-cli/references/short-drama/prompt/seedance-prompting.md +64 -0
  76. package/skills/using-vvicat-ai-short-studio-cli/references/short-drama/prompt-skill-index.md +2 -2
  77. package/skills/using-vvicat-ai-short-studio-cli/references/short-drama/recreation/recreation-qa.md +87 -0
  78. package/skills/using-vvicat-ai-short-studio-cli/references/short-drama/review/review-drama-script.md +2 -2
  79. package/skills/using-vvicat-ai-short-studio-cli/references/short-drama/review/review-drama-shots.md +6 -6
  80. package/skills/using-vvicat-ai-short-studio-cli/references/short-drama/storyboard/build-drama-storyboard.md +1 -3
  81. package/skills/using-vvicat-ai-short-studio-cli/references/short-drama/storyboard/design-fight-video.md +0 -4
  82. package/skills/using-vvicat-ai-short-studio-cli/references/short-drama/storyboard/direct-blender-previz/direction-contract.md +1 -1
  83. package/skills/using-vvicat-ai-short-studio-cli/references/short-drama/storyboard/direct-blender-previz.md +1 -1
  84. package/skills/using-vvicat-ai-short-studio-cli/references/short-drama/storyboard/generate-blender-previz.md +5 -10
  85. package/skills/using-vvicat-ai-short-studio-cli/references/short-drama/storyboard/plan-drama-production.md +2 -6
  86. package/skills/using-vvicat-ai-short-studio-cli/references/short-drama/storyboard/plan-shot-continuity.md +4 -4
  87. package/skills/using-vvicat-ai-short-studio-cli/references/short-drama/usage-guide.md +28 -70
  88. package/skills/using-vvicat-ai-short-studio-cli/references/hypit/SKILL.md +0 -447
  89. package/skills/using-vvicat-ai-short-studio-cli/references/hypit/creation/brief.md +0 -93
  90. package/skills/using-vvicat-ai-short-studio-cli/references/hypit/creation/project-files.md +0 -246
  91. package/skills/using-vvicat-ai-short-studio-cli/references/hypit/creation/reference-video.md +0 -266
  92. package/skills/using-vvicat-ai-short-studio-cli/references/hypit/creation/script-and-time.md +0 -253
  93. package/skills/using-vvicat-ai-short-studio-cli/references/hypit/creation/transformations.md +0 -141
  94. package/skills/using-vvicat-ai-short-studio-cli/references/hypit/environment/distribution.md +0 -176
  95. package/skills/using-vvicat-ai-short-studio-cli/references/hypit/environment/local-tools.md +0 -411
  96. package/skills/using-vvicat-ai-short-studio-cli/references/hypit/environment/model-and-provider.md +0 -275
  97. package/skills/using-vvicat-ai-short-studio-cli/references/hypit/environment/profile.md +0 -286
  98. package/skills/using-vvicat-ai-short-studio-cli/references/hypit/playbooks/craft/asset-and-shot-preproduction.md +0 -193
  99. package/skills/using-vvicat-ai-short-studio-cli/references/hypit/playbooks/craft/b-roll.md +0 -94
  100. package/skills/using-vvicat-ai-short-studio-cli/references/hypit/playbooks/craft/caption-tracking.md +0 -125
  101. package/skills/using-vvicat-ai-short-studio-cli/references/hypit/playbooks/craft/captions.md +0 -252
  102. package/skills/using-vvicat-ai-short-studio-cli/references/hypit/playbooks/craft/examples/conversation-images.md +0 -218
  103. package/skills/using-vvicat-ai-short-studio-cli/references/hypit/playbooks/craft/examples/image-direction.md +0 -123
  104. package/skills/using-vvicat-ai-short-studio-cli/references/hypit/playbooks/craft/generated-dependencies.md +0 -38
  105. package/skills/using-vvicat-ai-short-studio-cli/references/hypit/playbooks/craft/graphic-compositions.md +0 -85
  106. package/skills/using-vvicat-ai-short-studio-cli/references/hypit/playbooks/craft/image-direction.md +0 -95
  107. package/skills/using-vvicat-ai-short-studio-cli/references/hypit/playbooks/craft/motion-graphics.md +0 -68
  108. package/skills/using-vvicat-ai-short-studio-cli/references/hypit/playbooks/craft/recreation-policy.md +0 -51
  109. package/skills/using-vvicat-ai-short-studio-cli/references/hypit/playbooks/craft/screen-demonstrations.md +0 -51
  110. package/skills/using-vvicat-ai-short-studio-cli/references/hypit/playbooks/craft/shot-production-contracts.md +0 -161
  111. package/skills/using-vvicat-ai-short-studio-cli/references/hypit/playbooks/craft/sound-mix.md +0 -76
  112. package/skills/using-vvicat-ai-short-studio-cli/references/hypit/playbooks/craft/video-direction.md +0 -433
  113. package/skills/using-vvicat-ai-short-studio-cli/references/hypit/playbooks/craft/voice-and-performance.md +0 -224
  114. package/skills/using-vvicat-ai-short-studio-cli/references/hypit/playbooks/craft/voice-direction.md +0 -99
  115. package/skills/using-vvicat-ai-short-studio-cli/references/hypit/playbooks/formats/narration-led-demo.md +0 -87
  116. package/skills/using-vvicat-ai-short-studio-cli/references/hypit/playbooks/formats/presenter-led-explainer.md +0 -160
  117. package/skills/using-vvicat-ai-short-studio-cli/references/hypit/playbooks/formats/ranking-listicle.md +0 -71
  118. package/skills/using-vvicat-ai-short-studio-cli/references/hypit/playbooks/formats/reference-recreation-canvas.md +0 -19
  119. package/skills/using-vvicat-ai-short-studio-cli/references/hypit/playbooks/formats/short-drama-workflow.md +0 -165
  120. package/skills/using-vvicat-ai-short-studio-cli/references/hypit/playbooks/formats/short-drama.md +0 -88
  121. package/skills/using-vvicat-ai-short-studio-cli/references/hypit/playbooks/formats/street-interview.md +0 -92
  122. package/skills/using-vvicat-ai-short-studio-cli/references/hypit/playbooks/formats/talking-head.md +0 -82
  123. package/skills/using-vvicat-ai-short-studio-cli/references/hypit/playbooks/formats/two-person-podcast.md +0 -101
  124. package/skills/using-vvicat-ai-short-studio-cli/references/hypit/playbooks/index.md +0 -77
  125. package/skills/using-vvicat-ai-short-studio-cli/references/hypit/production/audio-presentation.md +0 -49
  126. package/skills/using-vvicat-ai-short-studio-cli/references/hypit/production/authoring.md +0 -228
  127. package/skills/using-vvicat-ai-short-studio-cli/references/hypit/production/browser-capture.md +0 -112
  128. package/skills/using-vvicat-ai-short-studio-cli/references/hypit/production/builds.md +0 -330
  129. package/skills/using-vvicat-ai-short-studio-cli/references/hypit/production/caption-authoring.md +0 -156
  130. package/skills/using-vvicat-ai-short-studio-cli/references/hypit/production/caption-presentation.md +0 -66
  131. package/skills/using-vvicat-ai-short-studio-cli/references/hypit/production/component-design.md +0 -217
  132. package/skills/using-vvicat-ai-short-studio-cli/references/hypit/production/component-sharing.md +0 -150
  133. package/skills/using-vvicat-ai-short-studio-cli/references/hypit/production/component-visuals.md +0 -152
  134. package/skills/using-vvicat-ai-short-studio-cli/references/hypit/production/examples/detail.svrun +0 -8
  135. package/skills/using-vvicat-ai-short-studio-cli/references/hypit/production/examples/direction.svs +0 -14
  136. package/skills/using-vvicat-ai-short-studio-cli/references/hypit/production/examples/look.svs +0 -6
  137. package/skills/using-vvicat-ai-short-studio-cli/references/hypit/production/examples/material.svrun +0 -5
  138. package/skills/using-vvicat-ai-short-studio-cli/references/hypit/production/examples/production.svml +0 -73
  139. package/skills/using-vvicat-ai-short-studio-cli/references/hypit/production/examples/production.svrun +0 -5
  140. package/skills/using-vvicat-ai-short-studio-cli/references/hypit/production/examples/visuals.ts +0 -62
  141. package/skills/using-vvicat-ai-short-studio-cli/references/hypit/production/fonts-and-text.md +0 -138
  142. package/skills/using-vvicat-ai-short-studio-cli/references/hypit/production/image-operations.md +0 -102
  143. package/skills/using-vvicat-ai-short-studio-cli/references/hypit/production/index.md +0 -73
  144. package/skills/using-vvicat-ai-short-studio-cli/references/hypit/production/media-presentation.md +0 -111
  145. package/skills/using-vvicat-ai-short-studio-cli/references/hypit/production/media.md +0 -246
  146. package/skills/using-vvicat-ai-short-studio-cli/references/hypit/production/performance.md +0 -114
  147. package/skills/using-vvicat-ai-short-studio-cli/references/hypit/production/prompt-kits.md +0 -165
  148. package/skills/using-vvicat-ai-short-studio-cli/references/hypit/production/rendering.md +0 -141
  149. package/skills/using-vvicat-ai-short-studio-cli/references/hypit/production/review.md +0 -211
  150. package/skills/using-vvicat-ai-short-studio-cli/references/hypit/production/runs.md +0 -151
  151. package/skills/using-vvicat-ai-short-studio-cli/references/hypit/production/script-syntax.md +0 -236
  152. package/skills/using-vvicat-ai-short-studio-cli/references/hypit/production/snapshots.md +0 -90
  153. package/skills/using-vvicat-ai-short-studio-cli/references/hypit/production/sound.md +0 -75
  154. package/skills/using-vvicat-ai-short-studio-cli/references/hypit/production/source-syntax.md +0 -163
  155. package/skills/using-vvicat-ai-short-studio-cli/references/hypit/production/spatial.md +0 -189
  156. package/skills/using-vvicat-ai-short-studio-cli/references/hypit/production/studio-companions.md +0 -186
  157. package/skills/using-vvicat-ai-short-studio-cli/references/hypit/production/studio.md +0 -249
  158. package/skills/using-vvicat-ai-short-studio-cli/references/hypit/production/system.md +0 -169
  159. package/skills/using-vvicat-ai-short-studio-cli/references/hypit/production/timeline.md +0 -74
  160. package/skills/using-vvicat-ai-short-studio-cli/references/hypit/production/timing.md +0 -96
  161. package/skills/using-vvicat-ai-short-studio-cli/references/hypit/production/track-authoring.md +0 -215
  162. package/skills/using-vvicat-ai-short-studio-cli/references/hypit/production/tracks.md +0 -67
  163. package/skills/using-vvicat-ai-short-studio-cli/references/hypit/production/video-downloads.md +0 -47
  164. package/skills/using-vvicat-ai-short-studio-cli/references/hypit/production/vocabulary.md +0 -119
  165. package/skills/using-vvicat-ai-short-studio-cli/references/short-drama/README.md +0 -11
  166. package/skills/using-vvicat-ai-short-studio-cli/references/short-drama/comfly-provider.md +0 -12
  167. package/skills/using-vvicat-ai-short-studio-cli/references/short-drama/generation/seedance-community-skills.md +0 -23
  168. package/skills/using-vvicat-ai-short-studio-cli/references/short-drama/recreation/publish-drama-references/litterbox.md +0 -5
  169. package/skills/using-vvicat-ai-short-studio-cli/references/short-drama/runninghub-provider.md +0 -33
  170. package/skills/using-vvicat-ai-short-studio-cli/references/short-drama/starrouter-provider.md +0 -53
  171. package/skills/using-vvicat-ai-short-studio-cli/references/video-model-routing.md +0 -10
  172. package/skills/using-vvicat-ai-short-studio-cli/references/video-prompt-spec.md +0 -32
  173. /package/skills/using-vvicat-ai-short-studio-cli/references/{short-drama/assets/modules/humanizer → humanizer}/LICENSE +0 -0
  174. /package/skills/using-vvicat-ai-short-studio-cli/references/{hypit/playbooks/craft → short-drama/prompt}/examples/h3-base-prompt-writing.md +0 -0
  175. /package/skills/using-vvicat-ai-short-studio-cli/references/{hypit/playbooks/craft → short-drama/prompt}/examples/h3-reference-prompt-writing.md +0 -0
@@ -1,9 +1,412 @@
1
- # 自然化短剧剧本
1
+ ---
2
+ name: humanizer
3
+ description: |
4
+ Remove signs of AI-generated writing from text. Use when editing or reviewing
5
+ text to make it sound more natural and human-written. Based on Wikipedia's
6
+ comprehensive "Signs of AI writing" guide. Detects and fixes patterns including:
7
+ inflated symbolism, promotional language, superficial -ing analyses, vague
8
+ attributions, em dash overuse, rule of three, AI vocabulary words, passive
9
+ voice, negative parallelisms, and filler phrases.
10
+ license: MIT
11
+ metadata:
12
+ version: "2.9.1"
13
+ ---
2
14
 
3
- 读取已选剧本、brief、bible、outline、source-analysis,以及用户明确提供的语言样本。只处理表达与可表演性,不新增事实,不改人物身份、关系、关键动作、场次顺序、伏笔兑现、时长目标或已确认台词含义。忠实改编不得用“更自然”为理由改写原作事实。
15
+ # Humanizer: Remove AI Writing Patterns
4
16
 
5
- 逐场检查并修正:角色都说同一种话、台词解释动机、连续完整句缺少打断与潜台词、旁白重复画面、舞台说明像散文、机械三段式、泛滥的转折词和总结句、口号式金句、同义反复、刻意悬念、过量标题与括号。让每个角色的词汇、句长、回避方式和情绪泄露符合人物档案;保留必要停顿、沉默、抢话和未说完的话,但不为“人味”堆口头禅。
17
+ You are a writing editor that identifies and removes signs of AI-generated text to make writing sound more natural and human. This guide is based on Wikipedia's "Signs of AI writing" page, maintained by WikiProject AI Cleanup.
6
18
 
7
- 输出完整剧本新版本,不输出修改建议代替成稿。用 `project-store.mjs put-script` 写入下一个不可变版本并 `select-script`,禁止覆盖原稿;随后必须执行 `review-drama-script`。用户语言样本只作为风格参考,不能复制其中的个人信息或无关事实。
19
+ ## Your Task
8
20
 
9
- 本 Skill 是插件内的短剧专用 Humanizer,吸收通用 Humanizer 的反模板化原则,但不套用百科写作规则,也不依赖外部 Skill。
21
+ When given text to humanize:
22
+
23
+ 1. **Identify AI patterns** - Scan for the patterns listed below.
24
+ 2. **Preserve the information, not the shape** - Every claim in the original survives into the rewrite, but depth doesn't have to be uniform: compress the dull parts, dwell where a human would, and merge or split paragraphs freely. When keeping the information and mirroring the original's structure pull in different directions, the information wins.
25
+ 3. **Never invent facts** - The rewrite must not contain any fact, name, number, date, quote, or citation that isn't in the source text. Swapping a vague claim for a specific one is allowed only when the specific comes from the source or from the user; if a sentence needs real-world detail to work, ask for it or write the plain version without it. Opinions and reactions are voice, not facts: where PERSONALITY AND SOUL applies you may add stance, but never new factual claims. (In fiction, invented detail is the job. This rule governs everything else.)
26
+ 4. **Match the voice** - Fit the intended tone (formal, casual, technical). Add personality only when the content and the author's voice call for it (see PERSONALITY AND SOUL).
27
+
28
+ How you're invoked changes what you deliver (see Invocation Modes). The draft → audit → final loop itself is defined under Process and Output, below.
29
+
30
+ ## Voice Calibration
31
+
32
+ If the user provides a writing sample (their own previous writing), analyze it before rewriting:
33
+
34
+ 1. Read the sample first. Note its sentence lengths, vocabulary, paragraph openings, punctuation, recurring phrases, and transitions.
35
+ 2. Match those habits instead of merely deleting AI patterns. Do not upgrade casual words or regularize deliberate quirks.
36
+ 3. Without a sample, use the default behavior below.
37
+
38
+ A sample outranks this skill's style rules, including the em dash rule in §14: if the sample uses em dashes, keep them at roughly the sample's frequency. Matching the author beats scrubbing the tell.
39
+
40
+ ## PERSONALITY AND SOUL
41
+
42
+ Avoiding AI patterns is only half the job. Sterile, voiceless writing is just as obvious as slop. Good writing has a human behind it.
43
+
44
+ **Apply this section only when the content and the author's voice call for it** - blog posts, essays, opinion, personal writing. For encyclopedic, technical, legal, or reference text, neutral and plain *is* the correct human voice; don't inject opinions or first person there.
45
+
46
+ When voice is appropriate, avoid uniform sentence structures, bloodless neutrality, and perfect organization. Let the writer have opinions, uncertainty, mixed feelings, humor, asides, and uneven rhythm. Never add factual claims to create that personality.
47
+
48
+ ## CONTENT PATTERNS
49
+
50
+ ### 1. Undue Emphasis on Significance, Legacy, and Broader Trends
51
+
52
+ **Words to watch:** stands/serves as, is a testament/reminder, a vital/significant/crucial/pivotal/key role/moment, underscores/highlights its importance/significance, reflects broader, symbolizing its ongoing/enduring/lasting, contributing to the, setting the stage for, marking/shaping the, represents/marks a shift, key turning point, evolving landscape, focal point, indelible mark, deeply rooted
53
+ **Problem:** LLM writing puffs up importance by adding statements about how arbitrary aspects represent or contribute to a broader topic.
54
+ **Before:**
55
+ > The Statistical Institute of Catalonia was officially established in 1989, marking a pivotal moment in the evolution of regional statistics in Spain. This initiative was part of a broader movement across Spain to decentralize administrative functions and enhance regional governance.
56
+ **After:**
57
+ > The Statistical Institute of Catalonia was established in 1989, part of a wider decentralization of administrative functions in Spain.
58
+
59
+ ### 2. Undue Emphasis on Notability and Media Coverage
60
+
61
+ **Words to watch:** independent coverage, local/regional/national media outlets, written by a leading expert, active social media presence
62
+ **Problem:** LLMs hit readers over the head with claims of notability, often listing sources without context.
63
+ **Before:**
64
+ > Her views have been cited in The New York Times, BBC, Financial Times, and The Hindu. She maintains an active social media presence with over 500,000 followers.
65
+ **After:**
66
+ > Her views have been cited in The New York Times and the BBC.
67
+
68
+ (If the source gives real context for one citation, what she said and where, keep that one and drop the rest of the list. Don't invent the context to make the trimmed version sound better.)
69
+
70
+ ### 3. Superficial Analyses with -ing Endings
71
+
72
+ **Words to watch:** highlighting/underscoring/emphasizing..., ensuring..., reflecting/symbolizing..., contributing to..., cultivating/fostering..., encompassing..., showcasing...
73
+ **Problem:** AI chatbots tack present participle ("-ing") phrases onto sentences to add fake depth.
74
+ **Before:**
75
+ > The temple's color palette of blue, green, and gold resonates with the region's natural beauty, symbolizing Texas bluebonnets, the Gulf of Mexico, and the diverse Texan landscapes, reflecting the community's deep connection to the land.
76
+ **After:**
77
+ > The temple is painted blue, green, and gold, colors meant to evoke Texas bluebonnets and the Gulf of Mexico.
78
+
79
+ ### 4. Promotional and Advertisement-like Language
80
+
81
+ **Words to watch:** boasts a, vibrant, rich (figurative), profound, enhancing its, showcasing, exemplifies, commitment to, natural beauty, nestled, in the heart of, groundbreaking (figurative), renowned, breathtaking, must-visit, stunning
82
+ **Problem:** LLMs have serious problems keeping a neutral tone, especially for "cultural heritage" topics.
83
+ **Before:**
84
+ > Nestled within the breathtaking region of Gonder in Ethiopia, Alamata Raya Kobo stands as a vibrant town with a rich cultural heritage and stunning natural beauty.
85
+ **After:**
86
+ > Alamata Raya Kobo is a town in the Gonder region of Ethiopia.
87
+
88
+ ### 5. Vague Attributions and Weasel Words
89
+
90
+ **Words to watch:** Industry reports, Observers have cited, Experts argue, Some critics argue, several sources/publications (when few cited)
91
+ **Problem:** AI chatbots attribute opinions to vague authorities without specific sources.
92
+ **Before:**
93
+ > Due to its unique characteristics, the Haolai River is of interest to researchers and conservationists. Experts believe it plays a crucial role in the regional ecosystem.
94
+ **After:**
95
+ > Researchers and conservationists study the Haolai River for its unusual characteristics.
96
+
97
+ (If a real source exists, name it. Never invent one to make a sentence sound sourced; an unsupported claim gets cut, not decorated.)
98
+
99
+ ### 6. Outline-like "Challenges and Future Prospects" Sections
100
+
101
+ **Words to watch:** Despite its... faces several challenges..., Despite these challenges, Challenges and Legacy, Future Outlook
102
+ **Problem:** Many LLM-generated articles include formulaic "Challenges" sections.
103
+ **Before:**
104
+ > Despite its industrial prosperity, Korattur faces challenges typical of urban areas, including traffic congestion and water scarcity. Despite these challenges, with its strategic location and ongoing initiatives, Korattur continues to thrive as an integral part of Chennai's growth.
105
+ **After:**
106
+ > Korattur has recurring traffic congestion and water shortages.
107
+
108
+ (The specifics you'd want here, like when the congestion worsened or what the city did about it, come from sources or the user, not from the rewrite.)
109
+
110
+ ## LANGUAGE AND GRAMMAR PATTERNS
111
+
112
+ ### 7. Overused "AI Vocabulary" Words
113
+
114
+ **High-frequency AI words:** Actually, additionally, align with, crucial, delve, emphasizing, enduring, enhance, fostering, garner, highlight (verb), interplay, intricate/intricacies, key (adjective), landscape (abstract noun), pivotal, showcase, tapestry (abstract noun), testament, underscore (verb), valuable, vibrant
115
+ **Problem:** These words appear far more frequently in post-2023 text. They often co-occur.
116
+ **Before:**
117
+ > Additionally, a distinctive feature of Somali cuisine is the incorporation of camel meat. An enduring testament to Italian colonial influence is the widespread adoption of pasta in the local culinary landscape, showcasing how these dishes have integrated into the traditional diet.
118
+ **After:**
119
+ > Somali cuisine also includes camel meat, which is considered a delicacy. Pasta dishes, introduced during Italian colonization, remain common, especially in the south.
120
+
121
+ ### 8. Avoidance of "is"/"are" (Copula Avoidance)
122
+
123
+ **Words to watch:** serves as/stands as/marks/represents [a], boasts/features/offers [a]
124
+ **Problem:** LLMs substitute elaborate constructions for simple copulas.
125
+ **Before:**
126
+ > Gallery 825 serves as LAAA's exhibition space for contemporary art. The gallery features four separate spaces and boasts over 3,000 square feet.
127
+ **After:**
128
+ > Gallery 825 is LAAA's exhibition space for contemporary art. The gallery has four rooms totaling 3,000 square feet.
129
+
130
+ ### 9. Negative Parallelisms and Tailing Negations
131
+ **Problem:** Constructions like "Not only...but..." or "It's not just about..., it's..." are overused. So are clipped tailing-negation fragments such as "no guessing" or "no wasted motion" tacked onto the end of a sentence instead of written as a real clause.
132
+ **Before:**
133
+ > It's not just about the beat riding under the vocals; it's part of the aggression and atmosphere. It's not merely a song, it's a statement.
134
+ **After:**
135
+ > The heavy beat adds to the aggressive tone.
136
+ **Before (tailing negation):**
137
+ > The options come from the selected item, no guessing.
138
+ **After:**
139
+ > The options come from the selected item without forcing the user to guess.
140
+
141
+ ### 10. Rule of Three Overuse
142
+ **Problem:** LLMs force ideas into groups of three to appear comprehensive.
143
+ **Before:**
144
+ > The event features keynote sessions, panel discussions, and networking opportunities. Attendees can expect innovation, inspiration, and industry insights.
145
+ **After:**
146
+ > The event includes talks and panels. There's also time for informal networking between sessions.
147
+
148
+ ### 11. Elegant Variation (Synonym Cycling)
149
+ **Problem:** AI has repetition-penalty code causing excessive synonym substitution.
150
+ **Before:**
151
+ > The protagonist faces many challenges. The main character must overcome obstacles. The central figure eventually triumphs. The hero returns home.
152
+ **After:**
153
+ > The protagonist faces many challenges but eventually triumphs and returns home.
154
+
155
+ ### 12. False Ranges
156
+ **Problem:** LLMs use "from X to Y" constructions where X and Y aren't on a meaningful scale.
157
+ **Before:**
158
+ > Our journey through the universe has taken us from the singularity of the Big Bang to the grand cosmic web, from the birth and death of stars to the enigmatic dance of dark matter.
159
+ **After:**
160
+ > The book covers the Big Bang, star formation, and current theories about dark matter.
161
+
162
+ ### 13. Passive Voice and Subjectless Fragments
163
+ **Problem:** LLMs often hide the actor or drop the subject entirely with lines like "No configuration file needed" or "The results are preserved automatically." Rewrite these when active voice makes the sentence clearer and more direct.
164
+ **Before:**
165
+ > No configuration file needed. The results are preserved automatically.
166
+ **After:**
167
+ > You do not need a configuration file. The system preserves the results automatically.
168
+
169
+ ## STYLE PATTERNS
170
+
171
+ ### 14. Em Dashes (and En Dashes): Cut Them
172
+
173
+ **Rule:** The final rewrite contains no em dashes (—) or en dashes (–). The em dash is one of the most reliable AI tells, so treat this as a hard constraint, not a "use sparingly" preference. Replace each one, in rough order of preference: a period (start a new sentence), a comma (a tight aside), a colon (introducing an explanation), parentheses (a true aside), or restructure the sentence. Also catch spaced em dashes (` — `) and double hyphens (` -- `) used the same way.
174
+ **Before:**
175
+ > The term is primarily promoted by Dutch institutions—not by the people themselves. You don't say "Netherlands, Europe" as an address—yet this mislabeling continues—even in official documents.
176
+ **After:**
177
+ > The term is primarily promoted by Dutch institutions, not by the people themselves. You don't say "Netherlands, Europe" as an address, yet this mislabeling continues in official documents.
178
+ **Before:**
179
+ > The new policy — announced without warning — affects thousands of workers. The changes -- long overdue according to critics -- will take effect immediately.
180
+ **After:**
181
+ > The new policy, announced without warning, affects thousands of workers. The changes, long overdue according to critics, will take effect immediately.
182
+
183
+ Before returning the final rewrite, scan it for `—` and `–`. Any hit means the draft isn't done. One exception: a user-provided writing sample that uses em dashes overrides this rule (see Voice Calibration); match the sample's frequency instead of banning them.
184
+
185
+ ### 15. Overuse of Boldface
186
+ **Problem:** AI chatbots emphasize phrases in boldface mechanically.
187
+ **Before:**
188
+ > It blends **OKRs (Objectives and Key Results)**, **KPIs (Key Performance Indicators)**, and visual strategy tools such as the **Business Model Canvas (BMC)** and **Balanced Scorecard (BSC)**.
189
+ **After:**
190
+ > It blends OKRs, KPIs, and visual strategy tools like the Business Model Canvas and Balanced Scorecard.
191
+
192
+ ### 16. Inline-Header Vertical Lists
193
+ **Problem:** AI outputs lists where items start with bolded headers followed by colons.
194
+ **Before:**
195
+ > - **User Experience:** The user experience has been significantly improved with a new interface.
196
+ > - **Performance:** Performance has been enhanced through optimized algorithms.
197
+ > - **Security:** Security has been strengthened with end-to-end encryption.
198
+ **After:**
199
+ > The update improves the interface, speeds up load times through optimized algorithms, and adds end-to-end encryption.
200
+
201
+ ### 17. Title Case in Headings
202
+ **Problem:** AI chatbots capitalize all main words in headings.
203
+ **Before:**
204
+ > ## Strategic Negotiations And Global Partnerships
205
+ **After:**
206
+ > ## Strategic negotiations and global partnerships
207
+
208
+ ### 18. Emojis
209
+ **Problem:** AI chatbots often decorate headings or bullet points with emojis.
210
+ **Before:**
211
+ > 🚀 **Launch Phase:** The product launches in Q3
212
+ > 💡 **Key Insight:** Users prefer simplicity
213
+ > ✅ **Next Steps:** Schedule follow-up meeting
214
+ **After:**
215
+ > The product launches in Q3. User research showed a preference for simplicity. Next step: schedule a follow-up meeting.
216
+
217
+ ### 19. Curly Quotation Marks
218
+ **Problem:** ChatGPT uses curly quotes (“...”) instead of straight quotes ("...").
219
+ **Before:**
220
+ > He said “the project is on track” but others disagreed.
221
+ **After:**
222
+ > He said "the project is on track" but others disagreed.
223
+
224
+ ## COMMUNICATION PATTERNS
225
+
226
+ ### 20. Collaborative Communication Artifacts
227
+
228
+ **Words to watch:** I hope this helps, Of course!, Certainly!, You're absolutely right!, Would you like..., Want me to...?, Want me to give examples?, Should I continue?, let me know, here is a...
229
+ **Problem:** Text meant as chatbot correspondence gets pasted as content.
230
+ **Before:**
231
+ > Here is an overview of the French Revolution. I hope this helps! Let me know if you'd like me to expand on any section.
232
+ **After:**
233
+ > The French Revolution began in 1789 when financial crisis and food shortages led to widespread unrest.
234
+
235
+ ### 21. Knowledge-Cutoff Disclaimers and Speculative Gap-Filling
236
+
237
+ **Words to watch:** as of [date], Up to my last training update, While specific details are limited/scarce..., based on available information, not publicly available, maintains a low profile, keeps personal details private, prefers to stay out of the spotlight, likely [grew up/studied/began], it is believed that
238
+ **Problem:** Two related tells. (a) Older models leave hard knowledge-cutoff disclaimers in the text. (b) When a model can't find a source, it writes a paragraph *about* not finding one and then invents plausible filler to cover the gap. For a private person the guess almost always lands on the same stock phrases ("maintains a low profile," "keeps personal details private"), none of it sourced. Say what isn't known, or cut the sentence; don't dress a guess up as fact.
239
+ **Before (cutoff disclaimer):**
240
+ > While specific details about the company's founding are not extensively documented in readily available sources, it appears to have been established sometime in the 1990s.
241
+ **After:**
242
+ > The company's founding date is not documented in the available sources. (Or cut the sentence. State a date only if a source provides one.)
243
+ **Before (speculative gap-fill):**
244
+ > Information about her early life is not publicly available, suggesting she maintains a low profile and keeps personal details private. She likely grew up in a middle-class household, which shaped her later interest in education reform.
245
+ **After:**
246
+ > Her early life is not documented in the available sources. (Or omit the section.)
247
+
248
+ ### 22. Sycophantic/Servile Tone
249
+ **Problem:** Overly positive, people-pleasing language.
250
+ **Before:**
251
+ > Great question! You're absolutely right that this is a complex topic. That's an excellent point about the economic factors.
252
+ **After:**
253
+ > The economic factors you mentioned are relevant here.
254
+
255
+ ## FILLER AND HEDGING
256
+
257
+ ### 23. Filler Phrases
258
+
259
+ **Before → After:**
260
+ - "In order to achieve this goal" → "To achieve this"
261
+ - "Due to the fact that it was raining" → "Because it was raining"
262
+ - "At this point in time" → "Now"
263
+ - "In the event that you need help" → "If you need help"
264
+ - "The system has the ability to process" → "The system can process"
265
+ - "It is important to note that the data shows" → "The data shows"
266
+
267
+ ### 24. Excessive Hedging
268
+ **Problem:** Over-qualifying statements.
269
+ **Before:**
270
+ > It could potentially possibly be argued that the policy might have some effect on outcomes.
271
+ **After:**
272
+ > The policy may affect outcomes.
273
+
274
+ ### 25. Generic Positive Conclusions
275
+ **Problem:** Vague upbeat endings.
276
+ **Before:**
277
+ > The future looks bright for the company. Exciting times lie ahead as they continue their journey toward excellence. This represents a major step in the right direction.
278
+ **After:**
279
+ > (Cut the paragraph. End on the last concrete fact instead of a send-off. If the source states real plans, use those.)
280
+
281
+ ### 26. Hyphenated Word Pair Overuse
282
+
283
+ **Words to watch:** third-party, cross-functional, client-facing, data-driven, decision-making, well-known, high-quality, real-time, long-term, end-to-end
284
+ **Problem:** AI hyphenates these uniformly, including in predicate position (`the report is high-quality`). Humans hyphenate inconsistently — typically only when the compound is attributive (`a high-quality report`) and often dropping the hyphen otherwise (`the report is high quality`). Keep attributive-position hyphens; drop them when the compound follows the noun.
285
+ **Before:**
286
+ > The cross-functional team delivered a high-quality, data-driven report. The team is cross-functional, the report is high-quality, and the methodology is data-driven.
287
+ **After:**
288
+ > The cross-functional team delivered a high-quality, data-driven report. The team is cross functional, the report is high quality, and the methodology is data driven.
289
+
290
+ ### 27. Persuasive Authority Tropes
291
+
292
+ **Phrases to watch:** The real question is, at its core, in reality, what really matters, fundamentally, the deeper issue, the heart of the matter
293
+ **Problem:** LLMs use these phrases to pretend they are cutting through noise to some deeper truth, when the sentence that follows usually just restates an ordinary point with extra ceremony.
294
+ **Before:**
295
+ > The real question is whether teams can adapt. At its core, what really matters is organizational readiness.
296
+ **After:**
297
+ > The question is whether teams can adapt. That mostly depends on whether the organization is ready to change its habits.
298
+
299
+ ### 28. Signposting and Announcements
300
+
301
+ **Phrases to watch:** Let's dive in, let's explore, let's break this down, here's what you need to know, now let's look at, without further ado
302
+ **Problem:** LLMs announce what they are about to do instead of doing it. This meta-commentary slows the writing down and gives it a tutorial-script feel.
303
+ **Before:**
304
+ > Let's dive into how caching works in Next.js. Here's what you need to know.
305
+ **After:**
306
+ > Next.js caches data at multiple layers, including request memoization, the data cache, and the router cache.
307
+
308
+ ### 29. Fragmented Headers
309
+
310
+ **Signs to watch:** A heading followed by a one-line paragraph that simply restates the heading before the real content begins.
311
+ **Problem:** LLMs often add a generic sentence after a heading as a rhetorical warm-up. It usually adds nothing and makes the prose feel padded.
312
+ **Before:**
313
+ > ## Performance
314
+ >
315
+ > Speed matters.
316
+ >
317
+ > When users hit a slow page, they leave.
318
+ **After:**
319
+ > ## Performance
320
+ >
321
+ > When users hit a slow page, they leave.
322
+
323
+ ### 30. Diff-Anchored Writing
324
+ **Problem:** Documentation or comments written as if narrating a change rather than describing the thing as it is. Unless the document is inherently version-scoped (changelogs, release notes, migration guides), it should read coherently without knowing what changed in the last commit.
325
+ **Before:**
326
+ > This function was added to replace the previous approach of iterating through all items, which caused O(n²) performance.
327
+ **After:**
328
+ > This function uses a hash map for O(1) lookups, avoiding the O(n²) cost of naive iteration.
329
+
330
+ ### 31. Manufactured Punchlines and Staccato Drama
331
+ **Problem:** LLMs often make every sentence land like a quotable closer, then stack short declarative fragments to manufacture drama. A single short sentence for emphasis is fine; a run of them starts to sound engineered.
332
+ **Before:**
333
+ > Then AlphaEvolve arrived. It had no preference for symmetry. No aesthetic prior. No nostalgia for human taste. The old rules were gone.
334
+ **After:**
335
+ > AlphaEvolve changed the search because it did not favor symmetry or human-looking designs. That made some of the older assumptions less useful.
336
+
337
+ ### 32. Aphorism Formulas
338
+
339
+ **Words to watch:** X is the Y of Z, X becomes a trap, X is not a tool but a mirror, the language of, the currency of, the architecture of
340
+ **Problem:** LLMs turn ordinary claims into reusable aphorisms that sound profound without adding precision. Replace the formula with the concrete claim it is gesturing at.
341
+ **Before:**
342
+ > Symmetry is the language of trust. Efficiency becomes a trap when teams forget the human layer.
343
+ **After:**
344
+ > Symmetric layouts often feel more predictable to users. Teams can over-optimize workflows and miss how people actually use them.
345
+
346
+ ### 33. Conversational Rhetorical Openers
347
+
348
+ **Phrases to watch:** Honestly?, Look, Here's the thing, The thing is, Let's be honest, Real talk, when used as standalone hooks or fake-candid pauses before an ordinary point.
349
+ **Problem:** LLMs open with a fake-candid hook to manufacture intimacy before delivering a routine claim. The tell is the theatrical pause-and-reveal: a one-word question or aside, then the "real" answer. A person being honest usually just says the thing.
350
+ **Before:**
351
+ > Is it worth the price? Honestly? It depends on how often you'll use it.
352
+ **After:**
353
+ > Whether it's worth the price depends on how often you'll use it.
354
+
355
+ ## DETECTION GUIDANCE
356
+
357
+ ### What NOT to flag (false positives)
358
+
359
+ A clean human writer can hit several of the patterns above without any AI involvement. Before rewriting, sanity-check that you are not gutting legitimate prose. The following are *not* reliable indicators on their own:
360
+
361
+ - **Perfect grammar and consistent style.** Many writers are professionals or have been edited. Polish does not equal AI.
362
+ - **Mixed casual and formal registers.** This often signals a person in a technical field, a young writer, or someone with neurodivergent prose habits — not a chatbot.
363
+ - **"Bland" or "robotic" prose.** AI prose has *specific* tells. Generic dryness without those tells is just dry writing.
364
+ - **Formal or academic vocabulary.** AI overuses *specific* fancy words (see §7), not all fancy words. Don't flatten "ostensibly" or "constituent" just because they sound brainy.
365
+ - **Letter-style opening or closing on a comment.** Salutations and sign-offs predate ChatGPT by centuries.
366
+ - **Common transition words in isolation.** *Additionally*, *moreover*, *consequently* are AI-coded only when piled up. One *however* is not a tell.
367
+ - **Curly quotes alone.** macOS, Word, Google Docs, and most CMSes auto-curl by default. Curly quotes only count when stacked with other tells.
368
+ - **Em dashes alone.** Many editors and journalists use them often. Em dashes are evidence only when paired with formulaic sales-y rhythm.
369
+ - **One short emphatic sentence.** Humans use clipped sentences to land a point. Flag staccato drama only when several short fragments appear in a row and inflate the tone.
370
+ - **"Honestly" or "look" mid-sentence.** These are ordinary in casual writing. The tell is the standalone theatrical opener, not the word itself.
371
+ - **Unsourced claims.** Most of the web is unsourced. Lack of citations doesn't prove anything.
372
+ - **Correct, complex formatting.** Visual editors and templates produce clean output without any AI.
373
+ - **Secondhand text.** Do not rewrite watched phrases inside quotations, titles, proper names, or examples where the phrase is being discussed rather than used.
374
+
375
+ When in doubt, look for **clusters** of tells, not isolated ones. A single em dash means nothing; em dashes plus rule-of-three plus *vibrant tapestry* plus a "Conclusion" section is a confession.
376
+
377
+ ### Signs of human writing (preserve these)
378
+
379
+ When you see these, lean toward leaving the prose alone — they are evidence of a real person writing, and over-editing will destroy what makes the piece sound human:
380
+
381
+ - **Specific, unusual, hard-to-fabricate detail.** A real address. A weird quote. The phrase "the lawyer who used to work upstairs from my dentist." LLMs round off specifics; humans hoard them.
382
+ - **Mixed feelings and unresolved tension.** "I think this is mostly good, but it bothers me, and I can't fully explain why." LLMs default to clean takes.
383
+ - **Dated, era-bound references.** Slang, memes, or in-jokes that map to a specific year and subculture. Models lag by a year or more.
384
+ - **First-person editorial choices the writer can defend.** If the writer can explain *why* they made a particular cut or used a particular word, that's a strong human signal.
385
+ - **Variety in sentence length.** Real writing alternates short and long. AI writing tends toward an even, mid-length cadence.
386
+ - **Genuine asides, parentheticals, or self-corrections.** "(I keep wanting to say 'almost' here, but it really was certain.)" Models rarely interrupt themselves like this.
387
+ - **Edits made before November 30, 2022.** ChatGPT's public launch. Anything older than that is, with very rare exceptions, not AI-written.
388
+
389
+ ---
390
+
391
+ ## Invocation Modes
392
+
393
+ **Pasted text (default).** The user gives text in the conversation. Run the full loop below and deliver the draft, the audit bullets, and the final rewrite.
394
+
395
+ **File mode.** The user points at a file. Read it, run the draft → audit → final loop internally, then rewrite the file in place so it ends up containing only the final rewrite. Humanize the prose only: leave code blocks, frontmatter, data, and link targets untouched. In the conversation, report a short summary of what changed rather than pasting the whole rewrite back.
396
+
397
+ **Embedded mode.** Another task or agent is using this skill as one step of a larger job (a PR description, a commit message, a doc). Run the loop internally and output only the final text. No draft, no audit bullets, no summary. The caller wants prose, not ceremony.
398
+
399
+ ## Process and Output
400
+
401
+ 1. Read the input carefully and identify every instance of the patterns above.
402
+ 2. Write a **draft rewrite**. Check that it reads naturally aloud, varies sentence length, prefers specific details and simple constructions (is/are/has), and keeps the appropriate register.
403
+ 3. Ask two questions: **"What makes the below so obviously AI generated?"** and **"Does the rewrite state any fact, name, number, date, or citation that isn't in the source?"** Answer briefly. A fabrication is a defect even when it sounds more human than the vague original.
404
+ 4. Revise into a **final rewrite** that addresses them and contains no em or en dashes (see §14).
405
+
406
+ In pasted-text mode, deliver the draft, the brief "still-AI" bullets, the final rewrite, and (optionally) a short summary of changes. In file and embedded modes, run the same loop but deliver only what the mode calls for (see Invocation Modes).
407
+
408
+ ## Reference
409
+
410
+ This skill is based on [Wikipedia:Signs of AI writing](https://en.wikipedia.org/wiki/Wikipedia:Signs_of_AI_writing), maintained by WikiProject AI Cleanup. The patterns documented there come from observations of thousands of instances of AI-generated text on Wikipedia.
411
+
412
+ Key insight from Wikipedia: "LLMs use statistical algorithms to guess what should come next. The result tends toward the most statistically likely result that applies to the widest variety of cases."
@@ -42,7 +42,7 @@
42
42
 
43
43
  ## 音乐生成节点
44
44
 
45
- 背景音乐、主题曲和纯音乐使用 `music-generate`,默认模型为 `starrouter::suno_music`。节点只接收可选的 `prompt` 文本输入,输出 AUDIO;在画布设置中通过 `musicModel` 继承项目音乐模型,节点 config 的 `model` 可以按实时目录覆盖。`prompt` 必须有实际内容,`title`、`tags`、`lyrics` 和 `makeInstrumental` 按实时 schema 提交;`lyrics` 与 `makeInstrumental: true` 互斥。不要用 `tts` 节点生成音乐,也不要把 `audioModel` 的 TTS 模型写到 `musicModel`。
45
+ 背景音乐、主题曲和纯音乐使用 `music-generate`,默认模型为 `starrouter::suno_music`。节点只接收可选的 `prompt` 文本输入,输出 AUDIO;在画布设置中通过 `musicModel` 继承项目音乐模型,节点 config 的 `model` 可以按实时目录覆盖。`prompt` 必须有实际内容,`title`、`tags`、`lyrics` 和 `makeInstrumental` 按实时 schema 提交;这里的 `title` 是 Suno 歌曲标题参数,其他节点不得把节点标题写进 config,标题一律用 `canvas node update --title`;`lyrics` 与 `makeInstrumental: true` 互斥。不要用 `tts` 节点生成音乐,也不要把 `audioModel` 的 TTS 模型写到 `musicModel`。
46
46
 
47
47
  创建后先回读节点类型、config 和 AUDIO 输出端口,再运行并等待任务。成功后回读 selected 版本,确认媒体类型为 `audio/*`、时长和文件非空并实际试听;任务失败或提交回执未知时先查任务,不自动重复发起付费请求。
48
48
 
@@ -1,38 +1,24 @@
1
1
  ---
2
2
  name: short-drama
3
3
  version: 1.0.0
4
- description: 短剧全流程唯一入口。用于原创短剧的创建、恢复、分析、编剧、导演、资产、分镜、打斗与 Blender 白模、图片/视频/声音生成、审核、剪辑、交付和市场研究;根据项目状态渐进读取 references 模块,不暴露或猜选子 Skill。参考片复刻统一使用原生 Hypit 链路。
4
+ description: 短剧全流程唯一入口。用于原创短剧的创建、恢复、分析、编剧、导演、资产、分镜、打斗与 Blender 白模、图片/视频/声音生成、审核、剪辑、交付和市场研究;根据项目状态渐进读取 references 模块,不暴露或猜选子 Skill。参考片复刻使用 CLI 复刻画布节点。
5
5
  ---
6
6
 
7
7
  # 短剧制作主 Skill
8
8
 
9
- 本 Skill 是插件唯一入口。专业能力保存在 `references/` 模块中;不要凭模块名自行模仿,也不要一次性读取全部 reference。始终由无限画布的 `canvas workflow status`、Hypit 文档版本和 `references/module-map.json` 决定当前需要加载的最小模块集合。新项目不得创建固定功能节点。
9
+ 本 Skill 是插件唯一入口。专业能力保存在 `references/` 模块中;不要凭模块名自行模仿,也不要一次性读取全部 reference。始终由无限画布的 `canvas workflow status`、项目文档版本和本目录 `module-map.json` 决定当前需要加载的最小模块集合。新项目不得预搭旧版固定功能节点链。
10
10
 
11
11
  ## 固定控制循环
12
12
 
13
- 进入控制循环前先确认项目类型:有参考片需要复刻、克隆或仿拍,或现有项目的 `workflow.type=viral-recreation` 时,立即转到原生 `hypit recreate`,不执行本插件的校验、模块路由或制作命令。以下步骤只适用于原创短剧的新建、恢复、继续和修改:
13
+ 进入控制循环前先确认项目类型:有参考片需要复刻、克隆或仿拍,或现有项目的 `workflow.type=viral-recreation` 时,先读取 [参考片复刻知识问答](recreation/recreation-qa.md),按其中的画布节点执行,不运行本插件的原创短剧模块。以下步骤只适用于原创短剧的新建、恢复、继续和修改:
14
14
 
15
- 1. 确认 AI Short Studio 项目和分集画布。除非用户明确要求本地 Hypit 文件项目,不在本地创建第二套项目状态。
16
- 2. 新项目读取 [项目管理模块](operations/manage-drama-projects.md) 后初始化;AI Short Studio 项目先运行 `canvas documents` 和 `canvas workflow status`。只有用户明确要求本地 Hypit 文件项目时,才读取 `.short-drama/` 状态文件。
17
- 3. 运行:
18
-
19
- ```bash
20
- node "${CODEX_PLUGIN_ROOT:-${CLAUDE_PLUGIN_ROOT:-.}}/scripts/validate-project.mjs" <项目目录>
21
- node "${CODEX_PLUGIN_ROOT:-${CLAUDE_PLUGIN_ROOT:-.}}/scripts/workflow.mjs" status <项目目录>
22
- node "${CODEX_PLUGIN_ROOT:-${CLAUDE_PLUGIN_ROOT:-.}}/scripts/module-runs.mjs" required <项目目录> <当前阶段>
23
- ```
24
-
25
- 在 AI Short Studio 画布中,以上状态由 `ai-short-studio canvas workflow status --project <projectId> --json` 提供;不要用本地 `workflow.mjs` 状态覆盖远程画布状态。
26
-
27
- 4. 对返回的每个 `module_id`,从 `references/module-map.json` 读取唯一 `path`,完整读取该 reference 后执行其输入、输出、失败关闭和验收合同。
28
- 5. 文本由 Codex 生成;只有图片、视频、声音和媒体变换调用 Provider。所有提示词使用 `assets/modules/<module_id>/prompts/` 中登记的模板并通过 `render-prompt.mjs` 留痕。
29
- 6. 每个模块完成后登记实际输入与证据:
30
-
31
- ```bash
32
- node "${CODEX_PLUGIN_ROOT:-${CLAUDE_PLUGIN_ROOT:-.}}/scripts/module-runs.mjs" record <项目目录> <阶段> <module_id> <项目内证据...> --input <项目内输入...>
33
- ```
34
-
35
- 7. 运行 `workflow.mjs check`;只有当前及已完成阶段的文档、资产、审核和 module-run 全部有效时才 `advance`。修改上游事实时按状态机回退,不手改完成状态。
15
+ 1. 确认 AI Short Studio 项目和分集画布;不要在本地创建第二套项目状态。
16
+ 2. 新项目读取 [项目管理模块](operations/manage-drama-projects.md) 后初始化;AI Short Studio 项目先运行 `canvas documents` 和 `canvas workflow status`。
17
+ 3. 运行 `ai-short-studio canvas workflow status --project <projectId> --json`,以远程画布状态和文档版本作为唯一事实来源;不要用本地状态文件覆盖。
18
+ 4. 对返回的每个 `module_id`,从本目录 `module-map.json` 读取唯一 `path`(相对插件根目录),完整读取该 reference 后执行其输入、输出、失败关闭和验收合同。
19
+ 5. 文本由 Codex 生成;只有图片、视频、声音和媒体变换调用 Provider。所有提示词使用 `assets/modules/<module_id>/prompts/` 中登记的模板。
20
+ 6. 每个模块完成后,把实际输入、已选版本和验收证据保存为新的 `canvas.<module>.v1` 文档版本,并用 `--input` 记录上游版本;不得用本地状态文件或本地完成标记代替服务端版本。
21
+ 7. 运行 `canvas workflow status` 复核当前阶段;只有当前及已完成阶段的文档、资产、审核和选版证据全部有效时才进入下一阶段。修改上游事实时回退受影响阶段并追加新版本,不手改完成状态。
36
22
 
37
23
  ## 九阶段基础路由
38
24
 
@@ -46,7 +32,7 @@ description: 短剧全流程唯一入口。用于原创短剧的创建、恢复
46
32
  | asset-analysis | `plan-drama-assets` → `generate-drama-art-style`,存在人物时增加 `generate-character-profiles` |
47
33
  | asset-generation | `manage-drama-assets`,缺少可复用资产时按需增加人物、场景、道具生成与生成服务模块 |
48
34
  | production-plan | 按条件加入 `design-fight-video`,再执行结构化分镜、制作计划与连续性模块 |
49
- | media-production | 原创项目按镜头选择分镜图类型、视频提示词、镜头审核、视频生成和任务监控;复杂调度按需增加 Blender 白模,按声音策略增加音频模块 |
35
+ | media-production | 原创项目按镜头生成已选择的分镜图类型并完成逐镜审计,再写视频提示词、生成视频和监控任务;复杂调度按需增加 Blender 白模,按声音策略增加音频模块 |
50
36
  | editing | `remotion-best-practices` → `edit-drama-timeline` |
51
37
  | delivery | `edit-deliver-drama` |
52
38
 
@@ -70,15 +56,15 @@ description: 短剧全流程唯一入口。用于原创短剧的创建、恢复
70
56
 
71
57
  ### 参考片复刻
72
58
 
73
- 参考片复刻不属于短剧插件;请退出本流程,使用原生 Hypit 复刻链路。旧项目会被插件阻止并提示迁移。
59
+ 参考片复刻不属于短剧插件;请读取 [参考片复刻知识问答](recreation/recreation-qa.md),使用 CLI 复刻画布节点。旧项目会被插件阻止并提示迁移。
74
60
 
75
61
  ### 声音
76
62
 
77
- 声音以成片原生声轨为先,先做七维复听并记录真实失败区间,再从最终 selected 音轨生成、复核字幕。合格原声不得重复 TTS 或整轨覆盖。StarRouter 尚无已验证的 CosyVoice 音色设计/克隆与 MuseTalk 口型能力;需要这些能力才能满足人物音色或可见对白同步合同时,明确阻塞并请用户调整制作方案,不得调用旧厂商或伪造通过。普通旁白、画外音只有在用户确认 StarRouter 可用音色与参数,且现有声音合同可满足时才能生成并完成八维复听。只有明确需要时才生成或授权 BGM、OP、ED;生成曲必须绑定剧情、人物主题、情绪弧、剪辑节奏和时长创作简报,未声明音乐不得交付。
63
+ 声音以成片原生声轨为先,先做七维复听并记录真实失败区间,再从最终 selected 音轨生成、复核字幕。合格原声不得重复 TTS 或整轨覆盖。StarRouter 不提供 CosyVoice 音色设计/克隆;百炼 CosyVoice 由平台实时目录注册,只有 `canvas node types` 返回对应节点和音色能力、且用户确认后才能用于音色设计。MuseTalk 口型驱动没有已验证入口;需要这些能力才能满足人物音色或可见对白同步合同时,明确阻塞并请用户调整制作方案,不得调用旧厂商或伪造通过。普通旁白、画外音只有在用户确认实时可用音色与参数,且现有声音合同可满足时才能生成并完成八维复听。只有明确需要时才生成或授权 BGM、OP、ED;生成曲必须绑定剧情、人物主题、情绪弧、剪辑节奏和时长创作简报,未声明音乐不得交付。
78
64
 
79
65
  ### 市场与支持操作
80
66
 
81
- 市场模块只在用户明确要求或项目已绑定市场输入时加载。Dashboard、Provider 配置、资产管理、任务监控、媒体变换和故障恢复都是按需 reference,不是独立 Skill;MCP、CLI 与 Dashboard 工具入口保持可用。
67
+ 市场模块只在用户明确要求或项目已绑定市场输入时加载。Dashboard、Provider 配置、资产管理、任务监控、媒体变换和故障恢复都是按需 reference,不是独立 Skill;CLI 与画布 Dashboard 入口保持可用。
82
68
 
83
69
  ## 全局硬门禁
84
70
 
@@ -88,7 +74,7 @@ description: 短剧全流程唯一入口。用于原创短剧的创建、恢复
88
74
  - 所有剧情事实、人物关系、结局、逐字对白与旁白变更先获得用户确认。
89
75
  - 画风确认必须先形成包含 `renderingContract` 的完整 style 对象;人物、场景、道具、已选分镜板和视频提示词一律从该对象编译媒介与渲染规则。不得把真人写实当成通用默认,也不得让“设定板/故事板/电影感”等版式或质量词覆盖用户选择的二维、三维、定格、像素或自定义媒介。
90
76
  - 成年人物突出帅气、漂亮、飒、性张力、独立妆造和记忆点;儿童突出可爱并禁止成人化或性化。
91
- - 任何人物、场景、道具、分镜、视频、音频和剪辑产物都必须登记真实本地路径、版本、来源和 SHA-256;不得用占位文件通过门禁。
77
+ - 任何人物、场景、道具、分镜、视频、音频和剪辑产物都必须登记为画布资产版本,并记录来源、选版和 SHA-256;不得用占位文件通过门禁。
92
78
  - 图片分镜、白模、视频、原生声音、兜底配音和成片必须执行各自 reference 的完整多维审核。未观看/未听完不得批准。
93
79
  - 复刻必须遵守权利范围;结构迁移不等于获得原人物、声音、音乐、字幕、品牌或镜头素材的使用权。
94
80
  - Provider 失败不得静默切换、重复付费或隐式降低质量。用户允许兜底时也必须记录原因、范围和新确认。
@@ -96,10 +82,10 @@ description: 短剧全流程唯一入口。用于原创短剧的创建、恢复
96
82
 
97
83
  ## 凭证与恢复
98
84
 
99
- `module-runs.json` 的每条记录绑定 module ID、reference 路径与哈希、输入与哈希、输出证据与哈希、prompt-run 及完成时间。reference、输入、选版或证据变化后记录自动失效。旧 `skill-runs.json` 永不原地修改;`module-runs.mjs migrate` 只迁移来源快照和证据均可验证的记录,其余模块重新执行。
85
+ `canvas.<module>.v1` 文档版本绑定 module ID、reference 路径、上游选版和输入来源;`canvas document save` 只追加不可变版本并记录 `--input`。reference、输入或选版变化后,旧版本仍可追溯,但不得继续作为当前阶段依据;不手工改写历史版本,也不运行本地迁移脚本。
100
86
 
101
87
  遇到失败先读取 [恢复模块](operations/recover-drama-pipeline.md),保持失败关闭。需要更新分镜时生成新不可变版本,不覆盖旧版本;需要重新生成媒体时先检查已有请求、任务和收据,避免重复提交。
102
88
 
103
89
  ## 完成标准
104
90
 
105
- 只有 delivery 阶段的 module-run、时间线审片、技术 QC、字幕、声音、权利和交付清单全部有效,才可运行 `workflow.mjs complete` 并向用户报告完成。报告必须区分已完成、已审核、待确认、失败和允许兜底的部分。
91
+ 只有 delivery 阶段的文档版本、时间线审片、技术 QC、字幕、声音、权利和交付清单全部有效,并通过 `canvas workflow status` 确认阶段完成后,才可向用户报告完成。报告必须区分已完成、已审核、待确认、失败和允许兜底的部分。
@@ -2,10 +2,10 @@
2
2
 
3
3
  主流程使用 `agent_character_visual`:`character_profiles` 只传已确认档案,`candidate_count` 传每个形象的候选数;输出必须是严格 JSON `characters[].appearances[].descriptions[]`。无档案的单人物请求才使用 `character_create`;修改、重生成和参考图反推分别使用 `character_modify`、`character_regenerate`、`character_description_update` 或 `character_image_to_description`。
4
4
 
5
- 正式出图必须用 `character_asset_sheet` 把已确认人物描述与项目完整画风对象编译为完整角色视觉设定板:脸部特写、正/侧/背全身主视图以及有证据的服装、配饰、鞋履和表情辅助图。它不是普通三视图,因此不再路由独立三视图 Skill。设定板只是参考信息布局,不能把媒介默认为概念原画;`render-prompt.mjs` 会把完整画风对象编译为 renderingContract,出图必须执行其媒介、材质与成像规则。
5
+ 正式出图必须用 `character_asset_sheet` 把已确认人物描述与项目完整画风对象编译为完整角色视觉设定板:脸部特写、正/侧/背全身主视图以及有证据的服装、配饰、鞋履和表情辅助图。人物设定板提示词遵循 [通用提示词规范](../prompt/prompt-craft.md),只写资产身份、视图、材质、光线和不可改变项。它不是普通三视图,因此不再路由独立三视图 Skill。设定板只是参考信息布局,不能把媒介默认为概念原画;登记模板会把完整画风对象编译为 renderingContract,出图必须执行其媒介、材质与成像规则。
6
6
 
7
7
  出图前必须对人物档案执行 `assertCharacterReadyForVisuals`:未知年龄不得生成视觉资产;成年人物必须保留已批准的 `adult-charisma`、专属妆发、服装标识和全部记忆锚点;儿童只允许 `child-cuteness`,禁止成人妆容、成人化服装、性感姿态、身体曲线强调和性化镜头语言。魅力设计服从时代、职业、阶层、伤病和剧情状态,不能靠无据裸露或模板化网红脸替代人物设计。
8
8
 
9
- 对所有无依赖形象全量提交 `generate_image`;StarRouter 并发由 Hypit Runtime Endpoint 控制,其余请求排队;同一人物的派生造型只在基础形象选版后进入下一并发层。`prompt_document` 必须引用该人物所在的当前 `asset-plan`:`{kind:"asset-plan",episode_key,version_id,asset_key}`;实际 prompt 必须等于 `character_asset_sheet` 的已留痕输出。参考图仅传该人物已选主形象,避免串脸。提交前展示并确认模型、数量、尺寸、参考路径和制作计划,不计算具体价格。检查脸型、发型、体型、服装版本、时代与标志物;每个候选都用 `asset-ledger.mjs fetch|decode` 保存到本地人物目录并写入相同制作文档引用。新生成候选禁止直接 `asset-ledger select`,必须先按 `review-drama-shots` 的人物专项合同通过保真预览与必要的局部裁切完整检查,再由 `review-ledger.mjs review-character` 原子审核和选版。审核绑定当前画风 ID 与内容哈希,并按用户选定媒介记录实际可见证据;不得固定使用真人写实标准。
9
+ 每个无依赖形象使用独立的 `character-turnaround` 节点(由画布按人物视觉方案物化,必要时再单独创建),`count` 只控制候选数,多结果不能拼图;同一人物的派生造型只在基础形象选版后进入下一并发层。节点 `prompt` 端口连接该人物的 `character-visual-design` 输出(`character-resource → character-visual-design → character-turnaround`),config 只写实时 schema 允许的 `count`、`aspectRatio` 和模型参数,不手写 `config.prompt` 冒充人物设定。节点正文由默认模板 `character_asset_sheet` 按当前 `asset-plan` 选版、已确认档案和完整项目画风对象编译,实际提交的 prompt 必须等于该模板输出;不得改成单张立绘,也不得改用 `image-generate` 代替,后者不追加角色设定板合同。`reference` 端口只连该人物已选主形象,避免串脸。提交前展示并确认模型、数量、尺寸、参考路径和制作计划,不计算具体价格。每个候选保存为画布资产版本;检查脸型、发型、体型、服装版本、时代与标志物。新生成候选禁止直接选版,必须先按 `review-drama-shots` 的人物专项合同通过保真预览与必要的局部裁切完整检查,再写入审核记录并原子审核和选版。审核绑定当前画风 ID 与内容哈希,并按用户选定媒介记录实际可见证据;不得固定使用真人写实标准。
10
10
 
11
11
  使用 StarRouter 时明确从以下枚举取值:`model=gpt-image-2`;`resolution ∈ {1K,2K,4K}`;`aspect_ratio ∈ {1:1,16:9,9:16,4:3,3:4}`;`quality ∈ {auto,low,medium,high}`;`background ∈ {auto,opaque,transparent}`;`moderation ∈ {auto,low}`;`output_format ∈ {png,jpeg,webp}`;`n=1..4`。人物资产默认 `background=opaque`;需要后期抠图时才确认 `transparent`。
@@ -1,7 +1,7 @@
1
1
  # 生成人物档案
2
2
 
3
- 先运行 `node "${CODEX_PLUGIN_ROOT:-${CLAUDE_PLUGIN_ROOT:-.}}/scripts/character-profiles.mjs" context <项目目录>`,将输出 JSON 作为 `characters_lib_info`;它是动态本地人物上下文,不是枚举。`input` 使用当前确认的小说或剧本文本。再用 `render-prompt.mjs` 渲染 `../../assets/modules/generate-character-profiles/prompts/agent_character_profile.{zh,en}.txt`:`character_profile_enums` 与 `character_profile_reference_vocabulary` 会从插件本地 [受控词表](../controlled-vocabulary.json) 自动注入,不得由用户临时填写。
3
+ 先生成人物档案文档,将输出 JSON 作为 `characters_lib_info`;它是动态本地人物上下文,不是枚举。`input` 使用当前确认的小说或剧本文本。再用登记的 `agent_character_profile.{zh,en}.txt` 模板编译:`character_profile_enums` 与 `character_profile_reference_vocabulary` 会从 [受控词表](../controlled-vocabulary.json) 注入,不得由用户临时填写。
4
4
 
5
5
  先合并别名、第一人称映射和关系,禁止把仅被提及或不出镜的对象建成人物。`performance_bible`、`voice_identity` 只记录原文明示或跨场景重复的长期特征;一次性动作、情绪和特效不得固化。每个角色至少保留初始形象,只有换装、年龄或持续造型变化才新增版本。
6
6
 
7
- 输出必须符合 [人物档案 Schema](generate-character-profiles/character-profile-output.schema.json),且顶层只能有 `new_characters[]` 与 `updated_characters[]`。先运行 `node "${CODEX_PLUGIN_ROOT:-${CLAUDE_PLUGIN_ROOT:-.}}/scripts/character-profiles.mjs" validate <项目目录> <结果 JSON>`;向用户展示新增、合并与冲突结果并确认后,才运行 `apply` 写入 `assets/characters/profiles.json`。人物事实确认后交给 `generate-character-images`;用户未确认前不得生成图片。
7
+ 输出必须符合 [人物档案 Schema](generate-character-profiles/character-profile-output.schema.json),且顶层只能有 `new_characters[]` 与 `updated_characters[]`。先生成人物档案文档;向用户展示新增、合并与冲突结果并确认后,才保存为项目文档版本。人物事实确认后交给 `generate-character-images`;用户未确认前不得生成图片。
@@ -1,9 +1,9 @@
1
1
  # 生成短剧画风
2
2
 
3
- 读取 brief、bible、用户参考和项目当前 `creative.art_style`。未提出自定义画风时保留默认 `system-realistic`;提出明确画风时,由 Codex 使用 `../../assets/modules/generate-drama-art-style/prompts/config_generation.{zh,en}.txt` 生成 prompt、description、renderingContract 和 visualBible,不调用文本 Provider。
3
+ 读取 brief、bible、用户参考和项目当前 `creative.art_style`。未提出自定义画风时保留默认 `system-realistic`;提出明确画风时,由 Codex 使用 `../assets/modules/generate-drama-art-style/prompts/config_generation.{zh,en}.txt` 生成 prompt、description、renderingContract 和 visualBible,不调用文本 Provider。
4
4
 
5
- 画风必须统一人物、场景、道具、分镜与视频的媒介、色板、光线和运动语言,但不得写入具体人物身份、剧情动作或地点事实。画风确认阶段就必须形成 `renderingContract`,明确 `medium`、`rendering_method`、`surface_language`、`image_formation` 和 `forbidden_substitutions`;不得把真人写实设为隐式默认,也不得用“电影感”“高级感”代替媒介定义。需要预览时调用 `drama-generation-service`,展示并确认图片 Provider、模型、尺寸、参考图和候选数,不计算具体价格;预览保存到本地 `assets/other/`。
5
+ 画风必须统一人物、场景、道具、分镜与视频的媒介、色板、光线和运动语言,但不得写入具体人物身份、剧情动作或地点事实。画风确认阶段就必须形成 `renderingContract`,明确 `medium`、`rendering_method`、`surface_language`、`image_formation` 和 `forbidden_substitutions`;不得把真人写实设为隐式默认,也不得用“电影感”“高级感”代替媒介定义。下游人物、场景、道具、分镜和视频提示词按 [通用提示词规范](../prompt/prompt-craft.md) 编译该合同,不在模板外另写风格词。需要预览时调用 `drama-generation-service`,展示并确认图片 Provider、模型、尺寸、参考图和候选数,不计算具体价格;预览保存为画布资产版本(`other-` 前缀)。
6
6
 
7
- 用户确认后,用 `project-store.mjs update-project` 将完整配置写入 `.short-drama/project.json` 的 `creative.art_style`。随后必须使用 `art_style_decision.{zh,en}.txt` 输出确认合同,并以 `project-store.mjs put-document <项目> art-style <JSON>` 保存为 `.short-drama/art-style.json`;其 style 必须与项目配置完全相同,才能登记本 Skill。下游 `render-prompt.mjs` 必须接收完整 style 对象,统一把它编译成画风合同后再注入人物、场景、道具、已选分镜板和视频提示词;不得只传名称、分类或一行风格短语。
7
+ 用户确认后,用 `canvas settings set` 将完整配置写入项目 `creative.art_style`。随后必须使用 `art_style_decision.{zh,en}.txt` 输出确认合同,并以 `canvas document save` 保存为 `art-style.v1` 文档;其 style 必须与项目配置完全相同,才能登记本 Skill。下游提示词模板必须接收完整 style 对象,统一把它编译成画风合同后再注入人物、场景、道具、已选分镜板和视频提示词;不得只传名称、分类或一行风格短语。
8
8
 
9
9
  当用户确认的是自定义画风(而非系统画风)时,必须将完整 style 对象追加或更新到统一工作区根目录下的 `art-styles.custom.json`(或环境变量 `SHORT_DRAMA_CUSTOM_ART_STYLES_FILE` 指定的路径),确保下次可直接选择。只登记已确认且通过完整字段校验的画风;不得把草稿、未确认结果或系统画风复制进自定义库,也不得覆盖系统画风同 ID。修改已投入使用的画风必须标记受影响资产为 stale,不静默混用旧风格。
@@ -2,10 +2,10 @@
2
2
 
3
3
  用 `select_prop`,向 `input` 传目标文本、向 `props_lib_name` 传已有道具库。只保留可移动、不可替代且有跨场景或长期持有证据的实体;普通家具、餐具、服装、场景设施和一次性物件不建资产。无合格项时接受 `{"props": []}`。
4
4
 
5
- `props[].description` 只写材质、颜色、形状、结构、数量与装饰。参考图或修改指令引起描述变化时,用 `prop_description_update`;出图时用 `prop_generate` 把已确认描述与完整项目画风对象编译为主视图、正/侧/背视图和有证据的结构细节设定板,不在调用处另拼隐藏提示词。设定板版式不得覆盖 renderingContract 选定的媒介与材质语言。
5
+ `props[].description` 只写材质、颜色、形状、结构、数量与装饰。道具设定板提示词遵循 [通用提示词规范](../prompt/prompt-craft.md),只写道具身份、视图、材质工艺和不可改变结构。参考图或修改指令引起描述变化时,用 `prop_description_update`;出图时用 `prop_generate` 把已确认描述与完整项目画风对象编译为主视图、正/侧/背视图和有证据的结构细节设定板,不在调用处另拼隐藏提示词。设定板版式不得覆盖 renderingContract 选定的媒介与材质语言。
6
6
 
7
7
  同一道具存在多个状态版本时,先生成并选定基础状态;后续状态必须把该选版作为身份参考,只改变版本证据明确要求的开合、破损、内容物或使用痕迹,并保持尺寸等级、结构、材质、纹样和独特锚点。基础状态缺少可用选版时不生成派生状态;不得用人物或手部充当尺度参照。
8
8
 
9
- 对同一依赖层的全部道具全量提交 `generate_image`;StarRouter 并发由 Hypit Runtime Endpoint 控制,其余请求排队;派生状态只在基础状态选版后进入下一并发层。生成白底居中、单道具资产图,传 `{kind:"asset-plan",episode_key,version_id,asset_key}`,并向 `render-prompt.mjs` 传完整项目画风对象;实际 prompt 必须等于 `prop_generate` 的已留痕输出。提交前展示并确认模型、数量、尺寸、参考路径和制作计划,不计算具体价格。用 `asset-ledger.mjs fetch|decode` 保存候选并写入相同制作文档引用;检查结构、数量、纹样、损伤状态以及 renderingContract 的媒介、材质和成像规则。新生成候选不得直接选版,必须执行 `review-ledger.mjs review-visual-asset`,绑定当前画风 ID 与内容哈希后才可原子选版。
9
+ 同一依赖层的每个道具使用独立的 `prop-turnaround` 节点(由画布按道具设定物化,必要时再单独创建),`count` 只控制候选数;派生状态只在基础状态选版后进入下一并发层。节点 `prompt` 端口连接该道具的 `prop-resource` 输出,config 只写实时 schema 允许的 `count`、`aspectRatio` 和模型参数,不手写 `config.prompt` 冒充道具设定。节点正文由默认模板 `prop_generate` 按当前 `asset-plan` 选版和完整项目画风对象编译,实际提交的 prompt 必须等于该模板输出;服务端会追加白底、单道具、主视图加正/侧/背三视图的资产图合同,不得改用 `image-generate` 或重新手写背景与视图要求。提交前展示并确认模型、数量、尺寸、参考路径和制作计划,不计算具体价格。每个候选保存为画布资产版本;检查结构、数量、纹样、损伤状态以及 renderingContract 的媒介、材质和成像规则。新生成候选不得直接选版,必须写入审核记录,绑定当前画风 ID 与内容哈希后才可原子选版。
10
10
 
11
11
  StarRouter 图片枚举:`model=gpt-image-2`;`resolution ∈ {1K,2K,4K}`;`aspect_ratio ∈ {1:1,16:9,9:16,4:3,3:4}`;`quality ∈ {auto,low,medium,high}`;`background ∈ {auto,opaque,transparent}`;`moderation ∈ {auto,low}`;`output_format ∈ {png,jpeg,webp}`;`n=1..4`。