Not in the top 1,000 by GitHub stars

chengzhi-c/natural-talk

让 AI 像人一样说话、创作的规则集。Ruleset for making AI talk and write like a human.

History

Review weekly changes to SKILL.md before updating. Dates mark the weeks in which we observed each version; checks apply to the recorded commits.

View weekly changes

`SKILL.md`

Week of Nov 2, 2026 · Line counts unavailable

Before Not checked yet → After Local scripts

First observation → 4bf7a0bf823e

No earlier version was observed for comparison.

Compare on GitHub

View on GitHub Add to my skills

Verified commit 4bf7a0bf823e6c3f5f043ba84a8ea8fb732cddd7

106GitHub stars

8stars in 7 days

1skill in this repository

1commit in 30 days

Heat

6.2

Star growth, last 7 daysNot available
Star growth, last 24 hoursNot available
Rank movement, last 7 daysNot available
Commits, last 30 daysNot available
Accounts asking about it through MCP, last 7 daysNot available
Hot votes this weekNot available

Heat never changes the order by stars, and sponsorship never touches it. How heat works

Stars over time

30-day change not yet available: not enough history.

Individual skills in this repository

natural-talk /

How it works

natural-talk

When your agent uses it

自然人味交互与去 AI 腔。消除客服腔、谄媚评判、免责包装、强制平衡与假大空套话；提供专业、平视的同行交流姿态、文本防套路准绳与叙事创作工法（支持日常对话、代码沟通与深度小说创作）；学术公文营销文案不适用。

What it loads into context

Frontmatter

99 tokens

Always available to the agent

SKILL.md

3,559 tokens

When the skill is activated

Reference documents

20,515 tokens

Loaded on request

Token counts are estimates. Scripts run separately; they are not counted as reference context.

The playbook

模式判定（按序命中即定）
零号原则：交互姿态（同行平视）
核心规则（生成自查）
防误杀原则（与核心规则同等约束力）
清理模式
交付前核验
场景路由

Scripts executed separately

evals/check-eval.py
evals/run-eval.py
evals/turing-eval.py
scripts/_loader.py
scripts/audit-cleanup.py
scripts/check-retention.py
scripts/scan-mechanical.py
scripts/test-audit-cleanup.py
scripts/test-corpus-human.py
scripts/test-daily-gen.py
scripts/test-eval-harness.py
scripts/test-positive-examples.py
scripts/test-rules-local.py
scripts/test-scan-mechanical.py
scripts/test-skill-contract.py
scripts/test-turing-eval.py
scripts/verify_repo.py

Files

.gitattributesasset98 bytes
.gitignoreasset409 bytes
CONTRIBUTING.mddoc4,175 bytes
LICENSEasset1,082 bytes
README.en.mddoc8,035 bytes
README.mddoc6,977 bytes
SKILL.mdcore11,914 bytes
assets/natural-talk.pngasset20,906 bytes
evals/benchmark.jsonlasset8,776 bytes
evals/cases/SF-01.txtdoc913 bytes
evals/cases/SF-02.txtdoc1,158 bytes
evals/cases/SF-03.txtdoc1,309 bytes
evals/cases/SF-04.txtdoc659 bytes
evals/cases/SF-05.txtdoc578 bytes
evals/cases/SF-06.txtdoc549 bytes
evals/cases/SF-07.txtdoc307 bytes
evals/cases/SF-08.txtdoc298 bytes
evals/cases/SF-09.txtdoc799 bytes
evals/cases/SNF-01.txtdoc286 bytes
evals/cases/SNF-02.txtdoc310 bytes
evals/cases/SNF-03.txtdoc241 bytes
evals/cases/SNF-04.txtdoc280 bytes
evals/cases/SNF-05.txtdoc163 bytes
evals/cases/SNF-06.txtdoc301 bytes
evals/cases/SNF-07.txtdoc304 bytes
evals/cases/SNF-08.txtdoc231 bytes
evals/check-eval.pyruns19,748 bytes
evals/manifest.jsonasset25,894 bytes
evals/run-eval.pyruns11,537 bytes
evals/turing-eval.pyruns11,073 bytes
references/fiction.mddoc28,829 bytes
references/rules-full.mddoc19,069 bytes
scripts/_loader.pyruns590 bytes
scripts/audit-cleanup.pyruns19,607 bytes
scripts/check-retention.pyruns2,242 bytes
scripts/fixtures/fiction-sample.txtdoc940 bytes
scripts/scan-mechanical.pyruns83,569 bytes
scripts/test-audit-cleanup.pyruns10,668 bytes
scripts/test-corpus-human.pyruns3,544 bytes
scripts/test-daily-gen.pyruns3,568 bytes
scripts/test-eval-harness.pyruns20,720 bytes
scripts/test-positive-examples.pyruns12,869 bytes
scripts/test-rules-local.pyruns14,147 bytes
scripts/test-scan-mechanical.pyruns70,905 bytes
scripts/test-skill-contract.pyruns3,264 bytes
scripts/test-turing-eval.pyruns6,747 bytes
scripts/verify_repo.pyruns21,238 bytes

Read SKILL.md


# Natural Talk：自然人味交互与去 AI 腔

让大模型摆脱客服腔、假平衡与机械感，回归真诚平视的专业同行交流。

本页是蒸馏核心：生成模式直接执行的核心规则与防误杀原则。全量编号规范
（D/B/C/N 四层，清理模式的引用依据）在 `references/rules-full.md`；
叙事工法细节在 `references/fiction.md`；长文机械自查跑 `scripts/scan-mechanical.py`。

### 全局默认（所有模式共用）

- 拿不准是否命中 → 视为不命中，保持原样。
- 代码块、行内代码、URL、YAML frontmatter、表格行、引用块逐字保护，任何模式不改动；普通对话不构成修改代码注释的授权。
- 用户明确要求的格式、风格、结构照做（"按小节写""用表格""讲讲原理"）；D 层偏好与 C1 让位于用户显式指令。

View the complete source on GitHub

Skills included

natural-talk

Repository facts

License: MIT
Created: Aug 12, 2026
Last pushed: Nov 4, 2026
Forks: 5
Open issues: 0
Project activity: Active

Before you install

Local scripts

Automated signals, not a guarantee

Checked Nov 6, 2026

No broad tool permissions
No dangerous shell commands
14 network referencesin 4 files: README.en.md:107 README.md:101 CONTRIBUTING.md:79 and 1 more
No hidden or obfuscated text
No instruction overrides
17 executable scriptsin 17 files: evals/check-eval.py:1 evals/run-eval.py:1 evals/turing-eval.py:1 and 14 more
8 references to secretsin 5 files: evals/run-eval.py:169 evals/turing-eval.py:191 scripts/test-eval-harness.py:143 and 2 more
No requests to vote in Hot

Show all 39 signals by file

`CONTRIBUTING.md`

Network accessline 79

- 在 [GitHub](https://github.com/chengzhi-c/natural-talk) 开 Issue

`README.en.md`

Network accessline 107

# Or clone into Claude Code / Cursor / Codex / Antigravity skills directory git clone https://github.com/chengzhi-c/natural-talk.git ~/.claude/skills/natural-talk

Network accessline 108

git clone https://github.com/chengzhi-c/natural-talk.git ~/.claude/skills/natural-talk

Network accessline 120

RikkaHub / SillyTavern: import `natural-talk.zip` from [Releases](https://github.com/chengzhi-c/natural-talk/releases).

Network accessline 151

- [shuorenhua](https://github.com/MrGeDiao/shuorenhua): Chinese-first rewrite skill; thanks for its exploration and insights into information conservation, editing boundaries, and engineering evaluations.

Network accessline 152

- [lieflat-less-ai-tone](https://github.com/larashero3-dotcom/lieflat-less-ai-tone): Chinese de-AI-tone prompt research; thanks for its empirical comparative study and criteria for the text-layer negative whitelist.

Network accessline 156

Join the [LINUX DO](https://linux.do) community, "a new ideal type of community".

`README.md`

Network accessline 101

# 或 clone 至 Claude Code / Cursor / Codex / Antigravity 技能目录 git clone https://github.com/chengzhi-c/natural-talk.git ~/.claude/skills/natural-talk

Network accessline 102

git clone https://github.com/chengzhi-c/natural-talk.git ~/.claude/skills/natural-talk

Network accessline 113

RikkaHub / SillyTavern：导入 [Release](https://github.com/chengzhi-c/natural-talk/releases) 的 `natural-talk.zip`。

Network accessline 144

- [shuorenhua](https://github.com/MrGeDiao/shuorenhua)：中文优先的去 AI 味改写 skill，感谢其在信息守恒、编辑边界与工程化评测上的探索与启发。

Network accessline 145

- [lieflat-less-ai-tone](https://github.com/larashero3-dotcom/lieflat-less-ai-tone)：中文去 AI 腔提示词研究，感谢其在文本层反清单与实证对照研究中提供的判定依据。

Network accessline 149

欢迎加入 [LINUX DO](https://linux.do) 社区，一个「新的理想型社区」。

`evals/check-eval.py`

Executable scriptline 1
```
Executable file
```

`evals/run-eval.py`

Executable scriptline 1
```
Executable file
```
Access to secretsline 169
```
[Credential-related source omitted]
```
Access to secretsline 170
```
[Credential-related source omitted]
```

`evals/turing-eval.py`

Executable scriptline 1
```
Executable file
```
Access to secretsline 191
```
[Credential-related source omitted]
```
Access to secretsline 192
```
[Credential-related source omitted]
```

`scripts/_loader.py`

Executable scriptline 1
```
Executable file
```

`scripts/audit-cleanup.py`

Executable scriptline 1
```
Executable file
```

`scripts/check-retention.py`

Executable scriptline 1
```
Executable file
```

`scripts/scan-mechanical.py`

Executable scriptline 1
```
Executable file
```

`scripts/test-audit-cleanup.py`

Executable scriptline 1
```
Executable file
```

`scripts/test-corpus-human.py`

Executable scriptline 1
```
#!/usr/bin/env python3
```

`scripts/test-daily-gen.py`

Executable scriptline 1
```
#!/usr/bin/env python3
```

`scripts/test-eval-harness.py`

Executable scriptline 1
```
Executable file
```
Access to secretsline 143
```
[Credential-related source omitted]
```
Access to secretsline 310
```
[Credential-related source omitted]
```

`scripts/test-positive-examples.py`

Executable scriptline 1
```
Executable file
```

`scripts/test-rules-local.py`

Executable scriptline 1
```
#!/usr/bin/env python3
```

`scripts/test-scan-mechanical.py`

Executable scriptline 1
```
Executable file
```

Network accessline 78

https://example.com/说白了、说穿了、先看这里

Access to secretsline 903
```
[Credential-related source omitted]
```

`scripts/test-skill-contract.py`

Executable scriptline 1
```
#!/usr/bin/env python3
```
Access to secretsline 16
```
[Credential-related source omitted]
```

`scripts/test-turing-eval.py`

Executable scriptline 1
```
Executable file
```

`scripts/verify_repo.py`

Executable scriptline 1
```
#!/usr/bin/env python3
```

How the signals fit together

Facts from a static read of the files at the pinned commit. They describe what the code does, not why.

Signal confidence: 39 facts

Network access in prose of documentationCONTRIBUTING.md:79Low confidence
Network access in code of documentationREADME.en.md:107Low confidence
Network access in code of documentationREADME.en.md:108Low confidence
Network access in prose of documentationREADME.en.md:120Low confidence
Network access in prose of documentationREADME.en.md:151Low confidence
Network access in prose of documentationREADME.en.md:152Low confidence
Network access in prose of documentationREADME.en.md:156Low confidence
Network access in code of documentationREADME.md:101Low confidence
Network access in code of documentationREADME.md:102Low confidence
Network access in prose of documentationREADME.md:113Low confidence
Network access in prose of documentationREADME.md:144Low confidence
Network access in prose of documentationREADME.md:145Low confidence
Network access in prose of documentationREADME.md:149Low confidence
Executable script in prose of a file that runsevals/check-eval.py:1Low confidence
Executable script in prose of a file that runsevals/run-eval.py:1Low confidence
Access to secrets in code of a file that runs: a variable, placeholder or call, not a stored valueevals/run-eval.py:169Low confidence
Access to secrets in code of a file that runs: a variable, placeholder or call, not a stored valueevals/run-eval.py:170Low confidence
Executable script in prose of a file that runsevals/turing-eval.py:1Low confidence
Access to secrets in code of a file that runs: a variable, placeholder or call, not a stored valueevals/turing-eval.py:191Low confidence
Access to secrets in code of a file that runs: a variable, placeholder or call, not a stored valueevals/turing-eval.py:192Low confidence
Executable script in prose of a file that runsscripts/_loader.py:1Low confidence
Executable script in prose of a file that runsscripts/audit-cleanup.py:1Low confidence
Executable script in prose of a file that runsscripts/check-retention.py:1Low confidence
Executable script in prose of a file that runsscripts/scan-mechanical.py:1Low confidence
Executable script in prose of a file that runsscripts/test-audit-cleanup.py:1Low confidence
Executable script in code of a file that runsscripts/test-corpus-human.py:1High confidence
Executable script in code of a file that runsscripts/test-daily-gen.py:1High confidence
Executable script in prose of a file that runsscripts/test-eval-harness.py:1Low confidence
Access to secrets in code of a file that runs: a variable, placeholder or call, not a stored valuescripts/test-eval-harness.py:143Low confidence
Access to secrets in code of a file that runs: a variable, placeholder or call, not a stored valuescripts/test-eval-harness.py:310Low confidence
Executable script in a comment of a file that runsscripts/test-positive-examples.py:1Low confidence
Executable script in code of a file that runsscripts/test-rules-local.py:1High confidence
Executable script in prose of a file that runsscripts/test-scan-mechanical.py:1Low confidence
Network access in code of a file that runsscripts/test-scan-mechanical.py:78High confidence
Access to secrets in code of a file that runs: a variable, placeholder or call, not a stored valuescripts/test-scan-mechanical.py:903Low confidence
Executable script in code of a file that runsscripts/test-skill-contract.py:1High confidence
Access to secrets in code of a file that runs: a variable, placeholder or call, not a stored valuescripts/test-skill-contract.py:16Low confidence
Executable script in prose of a file that runsscripts/test-turing-eval.py:1Low confidence
Executable script in code of a file that runsscripts/verify_repo.py:1High confidence

Works with

Agent compatibility, from the files in this repository
Agent	Status	Evidence	Install to
Claude Code	Compatible	Standard SKILL.md SKILL.md	`.claude/skills/<skill-name>/`
Codex	Compatible	Standard SKILL.md SKILL.md	`.agents/skills/<skill-name>/`
Cursor	Compatible	Standard SKILL.md SKILL.md	`.cursor/skills/<skill-name>/`
Gemini CLI	Compatible	Standard SKILL.md SKILL.md	`.gemini/skills/<skill-name>/`
GitHub Copilot	Compatible	Standard SKILL.md SKILL.md	`.github/skills/<skill-name>/`
OpenCode	Compatible	Standard SKILL.md SKILL.md	`.opencode/skills/<skill-name>/`

Native The repository ships a manifest, folder or frontmatter field for this agent.
Compatible A standard SKILL.md that this agent can read.
Unknown No evidence either way.

Derived automatically from paths, manifests and frontmatter. Not tested on each agent.

chengzhi-c/natural-talk

History

SKILL.md

Share this page

Heat

Stars over time

Individual skills in this repository

How it works

natural-talk

Frontmatter

SKILL.md

Reference documents

The playbook

Scripts executed separately

Files

Skills included

Repository facts

Before you install

CONTRIBUTING.md

README.en.md

README.md

evals/check-eval.py

evals/run-eval.py

evals/turing-eval.py

scripts/_loader.py

scripts/audit-cleanup.py

scripts/check-retention.py

scripts/scan-mechanical.py

scripts/test-audit-cleanup.py

scripts/test-corpus-human.py

scripts/test-daily-gen.py

scripts/test-eval-harness.py

scripts/test-positive-examples.py

scripts/test-rules-local.py

scripts/test-scan-mechanical.py

scripts/test-skill-contract.py

scripts/test-turing-eval.py

scripts/verify_repo.py

How the signals fit together

Works with

You might also use

0xNyk/council-of-high-intelligence

zenstory-ai/drama-skills

oil-oil/oil-motion

Astro-Han/karpathy-llm-wiki

wuji-labs/nopua

OneWave-AI/claude-skills

`SKILL.md`

`CONTRIBUTING.md`

`README.en.md`

`README.md`

`evals/check-eval.py`

`evals/run-eval.py`

`evals/turing-eval.py`

`scripts/_loader.py`

`scripts/audit-cleanup.py`

`scripts/check-retention.py`

`scripts/scan-mechanical.py`

`scripts/test-audit-cleanup.py`

`scripts/test-corpus-human.py`

`scripts/test-daily-gen.py`

`scripts/test-eval-harness.py`

`scripts/test-positive-examples.py`

`scripts/test-rules-local.py`

`scripts/test-scan-mechanical.py`

`scripts/test-skill-contract.py`

`scripts/test-turing-eval.py`

`scripts/verify_repo.py`