Story Multi-Perspective Review
zenstory-ai/oh-story-claudecode
Reviews Chinese web-novel text with several reviewer agents in parallel, falling back to a single-agent pass, and reports structure, character, prose and setting problems with fixes.
审核一段/一篇小说(或任意叙事性文本)的「AI 创作度」——判断它有多大可能由 AI 生成、或 AI 参与了多少。当用户想知道一段文字是不是 AI 写的、AI 味重不重、是人工还是 AI、AI 辅助了多少、帮忙鉴别/检测/审稿 AI 生成内容、给稿件的「人味/机味」打分时,都使用本技能。触发词包括但不限于:AI 创作度、AI 味、是不是 AI 写的、AI 生成检测、人工还是 AI、鉴别…
$ npx skills add Xiaoyangy/novel-studio --skill ai-novel-audit -a claude-codeProject install by default; add -g for ~/.claude/skills/.
$ gh skill install Xiaoyangy/novel-studio ai-novel-audit --agent claude-codeProject scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).
$ git clone --depth 1 https://github.com/Xiaoyangy/novel-studio.git skills-src && mkdir -p .claude/skills && cp -r skills-src/skills/review .claude/skills/ai-novel-audit && rm -rf skills-srcUse ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.
Claude Code skills documentation · loads skills from .claude/skills/
Install the "ai-novel-audit" agent skill from https://github.com/Xiaoyangy/novel-studio/tree/main/skills/review into .claude/skills/ai-novel-audit/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "ai-novel-audit", then confirm the skill loads.Claude Code copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$skill-installer install https://github.com/Xiaoyangy/novel-studio/tree/main/skills/reviewType this inside Codex. $skill-installer <name> installs a curated skill from openai/skills. The installer writes to $CODEX_HOME/skills (default ~/.codex/skills). Restart Codex if the skill does not show up.
$ npx skills add Xiaoyangy/novel-studio --skill ai-novel-audit -a codexProject install goes to .agents/skills/; add -g for ~/.codex/skills/.
$ gh skill install Xiaoyangy/novel-studio ai-novel-audit --agent codexProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/Xiaoyangy/novel-studio.git skills-src && mkdir -p .agents/skills && cp -r skills-src/skills/review .agents/skills/ai-novel-audit && rm -rf skills-srcUse ~/.agents/skills/ instead of .agents/skills for a personal install.
Codex skills documentation · loads skills from .agents/skills/
Install the "ai-novel-audit" agent skill from https://github.com/Xiaoyangy/novel-studio/tree/main/skills/review into .agents/skills/ai-novel-audit/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "ai-novel-audit", then confirm the skill loads.Codex copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add Xiaoyangy/novel-studio --skill ai-novel-audit -a cursorProject install goes to .agents/skills/; add -g for ~/.cursor/skills/.
$ gh skill install Xiaoyangy/novel-studio ai-novel-audit --agent cursorProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/Xiaoyangy/novel-studio.git skills-src && mkdir -p .cursor/skills && cp -r skills-src/skills/review .cursor/skills/ai-novel-audit && rm -rf skills-srcUse ~/.cursor/skills/ instead of .cursor/skills for a personal install.
Cursor skills documentation · loads skills from .cursor/skills/, .agents/skills/, .claude/skills/, .codex/skills/
Install the "ai-novel-audit" agent skill from https://github.com/Xiaoyangy/novel-studio/tree/main/skills/review into .cursor/skills/ai-novel-audit/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "ai-novel-audit", then confirm the skill loads.Cursor copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ gemini skills install https://github.com/Xiaoyangy/novel-studio.git --path skills/review--scope user (default) or --scope workspace; --path is the subfolder of the repo that holds the skill; --consent skips the security confirmation prompt.
$ npx skills add Xiaoyangy/novel-studio --skill ai-novel-audit -a gemini-cliProject install goes to .agents/skills/; add -g for ~/.gemini/skills/.
$ gh skill install Xiaoyangy/novel-studio ai-novel-audit --agent gemini-cliProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/Xiaoyangy/novel-studio.git skills-src && mkdir -p .gemini/skills && cp -r skills-src/skills/review .gemini/skills/ai-novel-audit && rm -rf skills-srcUse ~/.gemini/skills/ instead of .gemini/skills for a personal install, then run /skills reload.
Gemini CLI skills documentation · loads skills from .gemini/skills/, .agents/skills/
Install the "ai-novel-audit" agent skill from https://github.com/Xiaoyangy/novel-studio/tree/main/skills/review into .gemini/skills/ai-novel-audit/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "ai-novel-audit", then confirm the skill loads.Gemini CLI copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ gh skill install Xiaoyangy/novel-studio ai-novel-auditInstalls for Copilot at project scope by default; add --scope user for a personal install. Preview a skill first with gh skill preview. Needs GitHub CLI 2.90.0 or later (public preview).
$ npx skills add Xiaoyangy/novel-studio --skill ai-novel-audit -a github-copilotProject install goes to .agents/skills/; add -g for ~/.copilot/skills/.
$ git clone --depth 1 https://github.com/Xiaoyangy/novel-studio.git skills-src && mkdir -p .github/skills && cp -r skills-src/skills/review .github/skills/ai-novel-audit && rm -rf skills-srcUse ~/.copilot/skills/ instead of .github/skills for a personal install. Commit .github/skills so cloud agent and code review can use it.
GitHub Copilot skills documentation · loads skills from .github/skills/, .claude/skills/, .agents/skills/
Install the "ai-novel-audit" agent skill from https://github.com/Xiaoyangy/novel-studio/tree/main/skills/review into .github/skills/ai-novel-audit/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "ai-novel-audit", then confirm the skill loads.GitHub Copilot copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add Xiaoyangy/novel-studio --skill ai-novel-audit -a opencodeOpenCode documents no install command of its own. Project install goes to .agents/skills/; add -g for ~/.config/opencode/skills/.
$ gh skill install Xiaoyangy/novel-studio ai-novel-audit --agent opencodeProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/Xiaoyangy/novel-studio.git skills-src && mkdir -p .opencode/skills && cp -r skills-src/skills/review .opencode/skills/ai-novel-audit && rm -rf skills-srcUse ~/.config/opencode/skills/ instead of .opencode/skills for a personal install.
OpenCode skills documentation · loads skills from .opencode/skills/, .claude/skills/, .agents/skills/
Install the "ai-novel-audit" agent skill from https://github.com/Xiaoyangy/novel-studio/tree/main/skills/review into .opencode/skills/ai-novel-audit/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "ai-novel-audit", then confirm the skill loads.OpenCode copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
ai-novel-audit审核一段/一篇小说(或任意叙事性文本)的「AI 创作度」——判断它有多大可能由 AI 生成、或 AI 参与了多少。当用户想知道一段文字是不是 AI 写的、AI 味重不重、是人工还是 AI、AI 辅助了多少、帮忙鉴别/检测/审稿 AI 生成内容、给稿件的「人味/机味」打分时,都使用本技能。触发词包括但不限于:AI 创作度、AI 味、是不是 AI 写的、AI 生成检测、人工还是 AI、鉴别…
AI Novel Audit is an agent skill from Xiaoyangy/novel-studio. 审核一段/一篇小说(或任意叙事性文本)的「AI 创作度」——判断它有多大可能由 AI 生成、或 AI 参与了多少。当用户想知道一段文字是不是 AI 写的、AI 味重不重、是人工还是 AI、AI 辅助了多少、帮忙鉴别/检测/审稿 AI 生成内容、给稿件的「人味/机味」打分时,都使用本技能。触发词包括但不限于:AI 创作度、AI 味、是不是 AI 写的、AI 生成检测、人工还是 AI、鉴别 AI、查稿、审稿、AIGC 检测、机器味、文风鉴定。哪怕用户没明说「AI 创作度」这五个字,只要意图是评估一段叙事文本的 AI 生成可能性,就用本技能。注意:本技能只产出「概率性信号 + 证据」,绝不输出武断的非黑即白判决。
Its SKILL.md is about 3.1k tokens, which your agent loads only when the skill is triggered. The skill folder holds 2 other files (for example `CONTEXT.md` and `context.json`).
It sits in Writing & Content, covering Creative writing and fiction. The repository describes itself as: 开源、本地优先的 AI 长篇小说创作引擎:多智能体世界推演、按弧规划、RAG 长程记忆、逐章审核与断点恢复 | Open-source AI novel generator for long-form fiction. The licence is Apache-2.0.
7 steps, taken from the step headings in SKILL.md.
Read from SKILL.md and the folder at commit ed04a10. It shows what the files ask for, not the result of running them.
Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.
From allowed-tools in the SKILL.md frontmatter.
Shell commands in SKILL.md call:
python3From the folder's file list and the shell code blocks in SKILL.md.
No URLs in SKILL.md.
From URLs in SKILL.md, links to its own repository left out.
Names no API keys, tokens, secrets or passwords.
From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.
AI Novel Audit loads about 3.1k tokens when it runs. Until then it costs about 81 tokens; SKILL.md has 600 words of instructions outside code blocks.
Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.
The automated check found no risky patterns in SKILL.md.
Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.
The full file from Xiaoyangy/novel-studio at commit ed04a10, republished under its Apache-2.0 licence (© Xiaoyangy). 600 words, ~3,097 tokens.
.claude/skills/ai-novel-audit/SKILL.md (or your agent's skills folder). This skill also uses 2 other files; get the full folder from GitHub.帮助评估一段叙事性文本(小说、故事、散文片段、网文章节等)有多大可能由 AI 生成,或 AI 参与了多少创作。产出一份带证据的概率性评估报告,而非一句「是/否」的判决。
任何 AI 文本检测都不可靠,本技能也不例外。 在写报告和与用户沟通时,必须始终守住这条底线,原因是:
因此:
仓库内规范入口是 quality/audit/scripts/ 与 quality/audit/references/。本 skill 目录只保留流程说明,审核脚本和参考资料只维护在 quality/audit/。通过 novel-studio skills export --to <dir> 导出时,CLI 会把这些审核资源装配进导出产物的 review/ 目录。
.txt/.md 直接读;.docx/.pdf 用对应技能提取纯文本)。对全文运行(脚本只依赖标准库,无需联网):
python3 quality/audit/scripts/aigc_value.py <文本文件路径> --target 4
python3 quality/audit/scripts/text_signals.py <文本文件路径>
# 需要结构化结果时加 --json
# 只有用户主动给出腾讯朱雀/AIGC 外部平台抽查值时才原样带入;不得代替用户调用平台
python3 quality/audit/scripts/text_signals.py <文本文件路径> --external-aigc 0.8252aigc_value.py 会给出本地自研 自研AIGC值(0-1 小数 + 百分比),也就是本文的最终 AI占比;引擎名为 codex-local-aigc-v4。它融合概率曲率、弱语言模型一致性、局部熵/TTR、风格计量、全局/局部语义、叙事动力、内容完整性和分片代理。text_signals.py 会给出句长突发度(CV)、用字多样度、套路措辞、重复长片段、标点习惯、本地与综合风险分。这些是客观可复算的风险证据,但不是作者身份判决。
审核不是为了让低分文本看起来像合格正文。降 AI 味回改必须先保证剧情、人物、场景和读者读感,再处理检测通过性。以下内容一律视为内容完整性问题,不能作为降 AI 味手段:
正常的票据残字、聊天断句、设备误识别可以写,但必须短、可读、能推进剧情或揭示规则。
novel-studio/deconstruction-library/review-calibration/high-quality-human-prose 下的样本被用户标定为高质量人工写作、AI率为0。审核时要用它们校准误判:封闭场景、规则推理、强对话和类型名词稳定复现,会让弱模型曲线/滑窗熵看起来很稳,但只要没有脏码、真重复、工程词泄漏和空泛概括,这类稳定不能直接等同于 AIGC 高风险。
本地 aigc_value.py 会输出 human_anchor。小说正文命中 narrative_scene 锚点时,只能对概率曲线、弱语言模型、局部熵、风格计量和分片代理做软降权,不能覆盖整章单段风险,也不能再把结果固定压到 4.80%;技术说明文的 technical_expository 锚点单独校准。无论哪类锚点,都不覆盖内容完整性、真重复、工程词泄漏和项目严格 <4% 的交付门槛。
如果用户提供腾讯朱雀或其他 AIGC 平台抽查结果,必须单独记录,不要把它和本地六维评分或 DeepSeek provider judge 混在一起。系统不得主动打开、登录、提交或绕过验证码调用这些平台,也不得要求用户逐章回报:
0.8252 这类 0-1 小数必须换算为 82.52%。82.52% 这类百分比原样记录。4% 或以上时,否决这一版正文并触发一次整章重渲染。4% 或以上,这条新证据再触发该 SHA 的一次整章重渲染。4% 的抽查值也只是补充证据,不替代任何自动门禁。<4% 的结论,同时给出至少 2 条正文证据、2 条修改方案、1 条对白方案、1 条作者声口方案和 2 条可沉淀规则。建议缺失时即使分数 <4% 也不通过;下一轮必须按建议重排场景、情绪因果和对白参与者,不能只换词。automated_hard 配置(运行时 block_until_external_retest=true):此时自动 detector/mode 是同哈希硬门禁,通过后载荷冻结。用户手工抽查永远不会自动升级为 automated_hard。这一步是 n-gram 滑窗抓不到的,必须单独跑。 LLM 起草长文时,常把同一段场景、同一通对话、同一组金句在文章不同位置整段复述一遍——量化脚本会把这当作 n-gram 重复,但发布平台(番茄/起点/晋江)会判为"大段落重复"。
python3 quality/audit/scripts/paragraph_dup.py <文本文件路径>
# 单独跑也行;也可加到发布前自检流程脚本输出三组:
已知"故意的母题回扣"(比如一个 Slogan 在不同场景的反复出现)不算重复——脚本的 12 字级滑窗会保留这些"母题",但段落级检测只判独立的、间隔远的复述。两者是互补的,不是冲突的。
带着脚本结果,通读样本,按下面六个维度逐一评估。每个维度给一个 0–5 的「AI 倾向分」(0=完全人味,5=强烈 AI 味),并摘录原文短句作为证据。详细的信号清单和正反例见:
quality/audit/references/signals-zh.mdquality/audit/references/aigc-detection-current-notes.mdquality/audit/references/signals-en.md,并套用语言无关的通用原则按「评分方法」把六维分数合成 AI 创作度区间(%) + 置信度。
按「报告模板」组织。证据要具体到句子。
若用户想降低 AI 味或改稿,基于发现给具体的改写方向(而非泛泛而谈)。
下面给每个维度的核心问题、AI 倾向、人类倾向。完整清单与例句在 quality/audit/references/signals-zh.md。
脚本给的是信号,不是结论。读法:
>35/100 视为必须回改。aigc_value.py 的核心交付门控分。v4 在概率/熵/风格代理之外加入语意困惑度和叙事动力,专查对白传送带、动作标签同构、POV 内在体验薄、流程语汇与情绪范围过平。它不是外部平台判定,但服务端会自动运行,不能靠报告手写绕过。01.md、02.md、03.md 朱雀报告均将 3k 左右章节作为单片段处理,本地对 1800-3600 可见字符采用单段代理。若单片段疑似占比约 100%、片段分不低于 50%,分片风险下限直接采用该片段分;叙事人工锚点只能软降权,不能把这类整段疑似 AI 压成通过。semantic_noise_char_soup,说明文本里有长串无语法、无信息增量的汉字块;稀有神怪名词连续堆叠也算字符汤。发布流程必须先删改这类内容,用真实剧情材料替代。paragraph_dup.py 给出。完全重复段落和高度相似段落几乎一定是 LLM 自我抄袭,人类写作极少出现(偶尔是刻意的对仗/排比,但通常会变换措辞)。建议阈值:min-len=20 段落、min-distance=100 字符、min-chars=15 汉字。content_lint.py 给出,覆盖“X个字:……”这类可机械验证的数词错误,也覆盖“薄荷糖和创可贴两个字”这类把两个商品/词组误写成两个字的表达。命中 count_mismatch / two_items_as_two_chars 时按内容质量硬失败处理,不能因 AI 味低分放行。content_lint.py 给出,命中 forced_simile / stock_simile 时优先检查是否只是装饰。若出现在开篇、章末或密集出现,发布前必须重写。content_lint.py 给出,命中 dangling_order_word 时说明“先”等顺序词缺少参照,需改成明确状态句或补足后续顺序;命中 abrupt_strong_event 时说明“忽然/突然+强动作”缺少声源、视线、动作链或规则触发,需补足转场;命中 unsupported_speech_claim 时说明角色声称的听见/看见/知道缺少上文证据;命中 opaque_memo_shorthand 时说明备忘录/纸条缩写省掉了关键对象;命中 unit_name_apposition / clipped_habit_sentence 时说明句子像提纲省略,需补足“的/在/经常/他”等正常口语连接;命中 clipped_summary_phrase 时说明信息复盘句像报告摘要,需改成角色当下判断。content_lint.py 给出,命中 state_clause_pile 时说明同一句堆叠多个静态说明或重复“还”,需拆句或改为动作承接。把脚本数字翻译成维度 1、2 的打分依据,并在报告里引用具体数字当证据。
六维各打 0–5 分(0=完全人味,5=强 AI 味),锚点:
合成 AI 创作度:默认六维等权,AI创作度% = (六维总分 / 30) × 100,四舍五入到 5% 档,并表述为**±10% 的区间**(例:「约 65%,区间 55–75%」)。若某几个维度证据特别强,可适当加权并说明理由。
判定置信度(高/中/低),依据:
给出参考分档(描述用,别当铁律):
| AI 创作度 | 含义 |
|---|---|
| 0–20% | 几乎纯人工创作 |
| 20–40% | 以人工为主,可能有轻度 AI 润色 |
| 40–60% | 人机混合,难以判定 |
| 60–80% | 大量 AI 生成痕迹(可能 AI 起草 + 人工修改) |
| 80–100% | 高度疑似纯 AI 生成 |
发布/交付硬闸门(用于当前短篇生产流程):
自研AIGC值 必须严格低于目标值(当前为 4%);服务端会自动对正文运行,不是只读审核报告。AI创作度 与当前精确 SHA 的 DeepSeek provider judge 裸正文判定必须严格 <4%,且 provider 建议完整。六维合计 必须 ≤ 1,且任一单项不应 > 1。段落重复 必须为 0。本地AI味风险分 必须 ≤ 35/100。>=4%,该版触发一次整章重渲染;替换 SHA 不等待人工复测。仅显式 automated_hard 自动 detector/mode 例外。按以下结构输出(可用 Markdown,但保持简洁、可读):
# 小说 AI 创作度审核报告
## 结论
- AI 创作度:约 X%(区间 A–B%) —— 「以人工为主 / 人机混合 / 大量 AI 痕迹 …」
- 置信度:高 / 中 / 低
- 一句话判断:[用「呈现出…一致的特征」「疑似」一类措辞,不下死判]
- 交付闸门:通过 / 不通过
## 量化信号(脚本)
- 自研AIGC值 / AI占比:0.xxxx / X%;引擎 codex-local-aigc-v4;是否严格 `<4%`
- 朱雀四维代理分:突发性 X;困惑度代理 X;结构指纹 X;跨段一致性 X;最高风险维度及 stats/signals
- 近年检测器代理层:弱语言模型一致性 X;局部熵/TTR X;风格计量 X;语义平滑 X;最高风险 stats/signals
- 句长突发度 CV:数值(解读)
- 套路措辞密度:每千字 X 次;偏高类别:……
- 本地AI味风险分:X/100;综合AI味风险分:X/100
- 腾讯朱雀AIGC抽查值:原始值 / 换算百分比 / 对应正文 SHA / 是否触发当前版返工(未提供则写“未抽查/未知”)
- 重复长片段:……
- 段落重复:完全重复段落 X;高度相似段落 X;重复长句 X
- 其他:……
## 维度评分
| 维度 | AI倾向分(0–5) | 关键证据(摘录原文) |
|---|---|---|
| 1 句法与节奏 | x | 「……」 |
| 2 措辞与陈词 | x | 「……」 |
| 3 情感与心理 | x | 「……」 |
| 4 细节与质感 | x | 「……」 |
| 5 比喻与意象 | x | 「……」 |
| 6 声音/连贯/风险 | x | 「……」 |
## 主要依据(3–5 条,带原文证据)
1. ……
2. ……
## 降 AI 味回改清单
- 必修 1:原句/段落 → 回改方向
- 必修 2:原句/段落 → 回改方向
- 禁止用“整体更自然”这类空话替代具体改法
## 反面信号(指向「这是人写的」的线索,务必也列)
- ……(刻意找证伪线索,避免确认偏误)
## 重要提示
- 本结论为概率性文风分析,不能单独作为 AI 生成的定论,尤其不可直接用于指控或处分。
- 若需高可信度判定,请结合写作过程稿、版本历史等来源证据。务必保留「反面信号」一节:主动去找「这其实是人写的」的证据,对冲确认偏误。一份只会找 AI 痕迹、从不自我证伪的报告是不可信的。
quality/audit/scripts/text_signals.py —— 量化信号脚本(句长突发度、套路措辞密度、重复片段、段落级/句子级真重复),工作流第 2、3 步运行。quality/audit/scripts/paragraph_dup.py —— 独立的段落级真重复检测(发布系统级别),工作流第 3 步运行;也可单独跑用于发布前自检。quality/audit/scripts/content_lint.py —— 内容逻辑硬检脚本,目前覆盖“X个字:……”等数词事实错误、“两个商品/词组误写成两个字”、 “像一根刺”这类别扭/库存明喻、“先停了”这类顺序词悬空、“忽然砸门”这类无铺垫突发强事件、“我听见你说话了”这类证据声明悬空、“别回号”这类备忘录缩写、“1703蒋牧/搬来两个月,电梯里……”这类提纲式省略、“两个确认/最便宜的坑”这类摘要腔判断、状态说明逗号堆叠、恐慌台词句号切平/条款文本句号硬切,以及便签三条、黑卡 ToS、空对仗童谣、“X得发Y”复现、重复骂点、猫眼视角和身体/影子方位等表层结构感问题;命中时先改正文。quality/audit/references/signals-zh.md —— 中文叙事文本的 AI 信号详细清单与正反例,逐维度评估时查阅;含"段落级真重复"专节。quality/audit/references/aigc-detection-current-notes.md —— 公开 AI 文本检测逻辑与 codex-local-aigc-v4 本地化规则。quality/audit/references/signals-en.md —— 英文/其他语言文本的对应清单与语言无关原则。© Xiaoyangy, Apache-2.0. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file
SKILL.md and 2 other files in skills/review of Xiaoyangy/novel-studio.
Open the folder on GitHubat commit ed04a10
AI Novel Audit next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.
| Skill | Stars | Used in | Tokens | Auto-check | Licence | Repo updated |
|---|---|---|---|---|---|---|
| AI Novel Audit this skillXiaoyangy/novel-studio | 131 | — | ~3.1k | Automated safety check: Pass | Apache-2.0 | |
| Story Multi-Perspective Reviewzenstory-ai/oh-story-claudecode | 7.4k | 3 repos | ~3k | Automated safety check: Pass | MIT | |
| Short Web Fiction Trend Scanzenstory-ai/oh-story-claudecode | 7.4k | 2 repos | ~1.2k | Automated safety check: Pass | MIT | |
| InkOS Creative HarnessNarcooo/inkos | 10k | 1 repos | ~1.1k | Automated safety check: Pass | AGPL-3.0 | |
| Novel Arteternityspring/shuohao-skills | 4.3k | — | ~1.1k | Automated safety check: Notes | Apache-2.0 | |
| SepiaNanako0129/sepia | 3.1k | — | ~3.6k | Automated safety check: Pass | MIT |
zenstory-ai/oh-story-claudecode
Reviews Chinese web-novel text with several reviewer agents in parallel, falling back to a single-agent pass, and reports structure, character, prose and setting problems with fixes.
zenstory-ai/oh-story-claudecode
Scans popular short web-fiction rankings on Chinese platforms such as Dianzhong and Heiyan to surface trending emotional hooks, themes and topic candidates with an expiry warning.
Narcooo/inkos
Drives long-form fiction, scripts, storyboards, interactive films and long-document translation through InkOS, with every change made by a typed action.
eternityspring/shuohao-skills
给 AI 短剧出美术设定集(场景 + 叙事道具):场景的设计意图、一致性锚点、光照时段变体、 空景提示词;道具的戏剧功能、状态变体、尺度参照、白底无手提示词。
Nanako0129/sepia
Make AI-generated writing read as human-written, in fiction and in professional prose.
zenstory-ai/oh-story-claudecode
Routes a Chinese web-novel writing request to the matching tool in a 13-skill toolbox, covers author habit memory, and can launch a local dashboard for browsing a project.
Xiaoyangy/novel-studio
长篇网文写作方法论与 novel-studio pipeline 适配器。用于把长篇开书、大纲、日更、续写、重写等需求整理为 pipeline 输入;在 novel-studio 内禁止直接生成正文。触发方式:/story-long-write、/写长篇、「帮我开书」「写大纲」「日更」「续写」「继续写」「修改第X章」「回炉」「重写第X章」。
Xiaoyangy/novel-studio
短篇网文写作方法论与 novel-studio pipeline 适配器。用于把短篇、盐言故事、拆文续写等需求整理为 pipeline 输入;在 novel-studio 内禁止直接生成正文。触发方式:/story-short-write、/写短篇、「帮我写一篇短篇」「写个盐言故事」。
Xiaoyangy/novel-studio
豆瓣阅读原创长篇专项方法论与 novel-studio pipeline 适配器。用于把 15W-30W 字原创长篇的主题、类型、开头、文字、标签、简介和提交自检要求整理为 pipeline 输入;在 novel-studio 内禁止直接生成正文。触发方式:/story-douban-long-write、/豆瓣长篇、/豆瓣写作、「写一部豆瓣阅读原创长篇」「按豆瓣要求写长篇」。
Xiaoyangy/novel-studio
短篇网文拆文。拆解爆款短篇小说(番茄短篇 / 故事会 / 知乎盐选 / 追妻 / 世情 / 重生 / 虐渣等通俗题材)的故事核、结构、情感线、反转设计、写作手法、共鸣层次。单一全量拆解管道:跑完 Stage 2-6 产出完整拆文报告,落盘到 deconstruction-library/{书名}/,下游 story-short-write 同时读拆文报告 + 情节节点 + 写作手法 + 原文…
Xiaoyangy/novel-studio
把各功能串成一条可恢复的流水线:(共创→)写作→评审→重写→交付,按阶段执行、断点续跑。触发:「一条龙写完整本书」「从头跑完整个流程」「中断后接着跑」,想要无人值守端到端、且能中断恢复时使用。
Xiaoyangy/novel-studio
用一句创作需求驱动 novel-studio 走可恢复 pipeline 完整创作(写作→评审→重写→导出)。触发:「帮我写本小说」「按这个设定写」「跑一次创作」,已有明确创作要求且希望无人值守端到端产出章节时使用。
Categories
审核一段/一篇小说(或任意叙事性文本)的「AI 创作度」——判断它有多大可能由 AI 生成、或 AI 参与了多少。当用户想知道一段文字是不是 AI 写的、AI 味重不重、是人工还是 AI、AI 辅助了多少、帮忙鉴别/检测/审稿 AI 生成内容、给稿件的「人味/机味」打分时,都使用本技能。触发词包括但不限于:AI 创作度、AI 味、是不是 AI 写的、AI 生成检测、人工还是 AI、鉴别…. AI Novel Audit is an agent skill from Xiaoyangy/novel-studio.
AI Novel Audit fits situations like: tasks that involve Creative writing and fiction.
Run `npx skills add Xiaoyangy/novel-studio --skill ai-novel-audit -a claude-code`. Or copy the skill folder (skills/review in Xiaoyangy/novel-studio) into .claude/skills/ai-novel-audit in your project. Claude Code loads it when a task matches its description.
Run `npx skills add Xiaoyangy/novel-studio --skill ai-novel-audit -a codex`. Or copy the skill folder (skills/review in Xiaoyangy/novel-studio) into .agents/skills/ai-novel-audit in your project. Codex loads it when a task matches its description.
Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add Xiaoyangy/novel-studio --skill ai-novel-audit -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/ai-novel-audit, .gemini/skills/ai-novel-audit, .github/skills/ai-novel-audit and .opencode/skills/ai-novel-audit in your project.
Going by SKILL.md and its folder, AI Novel Audit needs the command-line tools its instructions call (python3). Our summary lists: Python 3.
SKILL.md contains no URLs. Any network use would come from the scripts or tools the agent runs. This is read from the text; nothing was executed.
Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. Review the folder before installing.
AI Novel Audit is published under the Apache-2.0 licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.
About 3.1k tokens (SKILL.md is roughly 12k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full.
Skills that share tags, products or a category with AI Novel Audit: Story Multi-Perspective Review (zenstory-ai/oh-story-claudecode, 7.4k stars), Short Web Fiction Trend Scan (zenstory-ai/oh-story-claudecode, 7.4k stars), InkOS Creative Harness (Narcooo/inkos, 10k stars) and Novel Art (eternityspring/shuohao-skills, 4.3k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.
Xiaoyangy (a GitHub user) maintains it in Xiaoyangy/novel-studio, which has 131 GitHub stars. The repository holds 18 skills in this directory. The repository was last updated on September 23, 2026.
Source: Xiaoyangy/novel-studio on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.