← Back to Skills Marketplace
Novel Publish Ready
by
chn012cjus
· GitHub ↗
· v1.0.9
· MIT-0
232
Downloads
1
Stars
0
Active Installs
10
Versions
Install in OpenClaw
/install web-novel-publishing-readiness-and-quality-check-skill
Description
小说质量检查技能。触发关键词：检查正文、质量报告、违禁词、套路句、章节衔接、逻辑漏洞、自检、人写感、大纲、人设。执行最大算力深度推理五步链，每章必须跑freq_check.py词频扫描+逐行违禁词扫描双轨制，复核自检通过方可出报告。
Usage Guidance
This skill appears coherent and works locally: it reads a text file, detects forbidden phrases and repeated short Chinese words, and writes freq_result.txt in the script folder. Before installing or running: 1) confirm you are comfortable with the agent/model being handed the chapter text(s) (SKILL.md asks for cross‑chapter comparisons and multi-step reasoning), 2) note the SKILL.md promises a persistent 'dynamic high-frequency word library' but the provided script does not implement cross-run persistence—if you rely on that feature, ask the author for clarification or a revised implementation, 3) the script writes freq_result.txt adjacent to the script (it may overwrite previous runs), and 4) there is no networking or credential access in the code, so no obvious exfiltration risk in the provided files. If you need the cross-chapter or persistent-database behavior, request the missing implementation details and confirm where/how persistent data will be stored.
Capability Analysis
Type: OpenClaw Skill
Name: web-novel-publishing-readiness-and-quality-check-skill
Version: 1.0.9
The skill bundle is a specialized tool for auditing web novel drafts for quality, consistency, and 'AI-like' writing tropes. The SKILL.md provides detailed instructions for a multi-step editorial review, and the freq_check.py script is a straightforward text processor that identifies overused phrases and forbidden words without any network access, obfuscation, or unauthorized file system operations.
Capability Assessment
Purpose & Capability
Name/description, SKILL.md, and freq_check.py all focus on novel-quality checks (word-frequency and forbidden-word scanning). Most required behaviors are implemented locally. Minor mismatch: SKILL.md describes a '动态高频词库' that accumulates repeated phrases across chapters and '下次检查同步扫描', but freq_check.py does not implement a persistent shared database—it only writes a per-run freq_result.txt. Also SKILL.md expects cross-chapter 500‑word comparisons, but the script only processes a single file passed as argument.
Instruction Scope
SKILL.md instructs the agent to perform five-step deep reasoning and multi-step checks (including cross-chapter comparisons). Those are workflow rules for the model, not system actions. The only system operation the script performs is reading the provided text file and writing freq_result.txt in the script directory. The instructions do not request unrelated system files or credentials. The cross-chapter check implies the agent must obtain the previous chapter text; that is operational (requires supplying both files) but not a hidden or excessive access request.
Install Mechanism
No install spec; the skill is instruction-only with a small local Python script. Nothing is downloaded or installed automatically.
Credentials
The skill requests no environment variables, no credentials, and no config paths. The script runs locally on an input file and writes a single output file; this is proportionate to the stated purpose.
Persistence & Privilege
The skill does not request always:true and has no special persistence. It writes freq_result.txt in its script directory (per-run artifact), which is expected behavior for a local scanner.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install web-novel-publishing-readiness-and-quality-check-skill - After installation, invoke the skill by name or use
/web-novel-publishing-readiness-and-quality-check-skill - Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.9
Version 1.0.9 – Major framework and rule overhaul for quality checks
- 增加底层创作规则：禁止使用AI训练语料和成品句式，仅允许用字典词汇全新组合形成文本。
- 强化推理链执行：每次检查必须执行五步完整推理链，包括任务解析、多分支解法、全链路推理、交叉校验和交付物生成，严禁省略或简化任一步骤。
- 新增替代表达库建设原则和动态高频词库更新机制，要求高频重复词进入违禁库并持续滚动更新。
- 明确替换方案自检流程，所有修改建议必须先违禁词和逻辑校验双重通过，禁止出现用禁用词替换禁用词、语义偏移或无关感情色彩。
- 检查流程与输出格式保持原有细项，进一步补充了违禁内容清单、具体操作边界和风险提示
v1.0.8
Version 1.0.8
- 明确增加“比喻词”（如、像、如同、仿佛、好像）为禁用项，并在违禁词与词频检查中单独列出，不再仅作为比喻句泛指。
- 明确要求“比喻词”在全文禁止，词频规则同步更新，避免比喻句重复限制和“比喻词”禁用矛盾。
- 优化违禁词与模板化句子检查相关表述，提高检查项准确性和覆盖面。
- 不涉及代码调整，仅SKILL.md文档更新。
v1.0.7
## Version 1.0.7 Changelog
- 新增词频分析结果文件 `freq_result.txt`，用于加强词频穷尽扫描流程的合规性和可追溯性。
- 检查流程及工具依赖未变，既有文本规范和操作指引保持一致。
- 主要提升实际检测环节的操作闭环，便于发现章节内高频词及重复问题。
v1.0.6
**词频检查脚本freq_check.py集成，提升自动化与流程合规性：**
- 新增freq_check.py脚本，实现整章Python词频自动检查，强制每章执行。
- SKILL.md文档中显式加入“工具依赖”说明，要求使用freq_check.py，无需额外安装。
- 明确强调词频扫描“不得替代”违禁词人工逐条扫描，两者缺一不可。
- 检查流程第四步细化自动词频检测的适用范围与注意事项，提升合规性。
- 其他文本内容和检查流程保持一致，全面保障质量检查流程自动化和规范性。
v1.0.5
Minor update clarifying batch reporting and post-fix review requirements for quality check.
- 指定：批量检查角色命名时，发现地摊式称呼后须全文扫描，报告所有相关问题，不得只报单例。
- 新增：修正完毕后，要求重新进行违禁词和词频扫描，确认无新增问题方可报告完成，禁止假完成。
v1.0.4
No functional or content changes detected for version 1.0.4.
- No files were changed in this release.
- Behavior and documentation remain as in the previous version.
- No impact for current users.
v1.0.3
No functional or content changes. Documentation updated with an added命名规范（character naming rule）under“人设一致性检查”（character consistency check）：
- 明确要求有台词NPC须用正式名字，禁止用“刀疤脸”等地摊式称呼。
- 无台词背景角色可用描述性称呼。
No other substantive changes detected in this version.
v1.0.2
**Notice:** This version clarifies and strengthens requirements for web novel quality checks, especially around duplicate word scanning.
- Added a clear warning that step 4 (词频规则检查) cannot replace step 2 (违禁词扫描); both must be executed independently.
- Explained that the word frequency script only catches repeated words and cannot detect single-instance banned/formulaic phrases.
- Stressed that running only the word frequency script is not sufficient for a "pass".
- No other content or workflow changes.
v1.0.1
Minor update to core检查原则，修正对输出建议的更严格要求。
- 新增“禁用词优先级”由三级提升为四级铁律，增加第5条：修正建议本身也需通过违禁词扫描，杜绝“用禁用词A替换禁用词B”的情况。
- 其它检查流程、规则和报告模板保持不变。
v1.0.0
- Initial release of the novel quality check skill.
- Provides comprehensive multi-step quality assessment for novel manuscripts, covering prohibited word scan, cliché detection, logic consistency, chapter transition, character/world consistency, outline alignment, pacing, repetition, terminology, and self-review.
- Emphasizes “human-like writing” and strictly enforces banned words and template replacement principles.
- Includes detailed output format for standardized quality reports.
Metadata
Frequently Asked Questions
What is Novel Publish Ready?
小说质量检查技能。触发关键词：检查正文、质量报告、违禁词、套路句、章节衔接、逻辑漏洞、自检、人写感、大纲、人设。执行最大算力深度推理五步链，每章必须跑freq_check.py词频扫描+逐行违禁词扫描双轨制，复核自检通过方可出报告。 It is an AI Agent Skill for Claude Code / OpenClaw, with 232 downloads so far.
How do I install Novel Publish Ready?
Run "/install web-novel-publishing-readiness-and-quality-check-skill" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is Novel Publish Ready free?
Yes, Novel Publish Ready is completely free, licensed under MIT-0. You can download, install and use it at no cost.
Which platforms does Novel Publish Ready support?
Novel Publish Ready is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created Novel Publish Ready?
It is built and maintained by chn012cjus (@chn012cjus); the current version is v1.0.9.
More Skills