- name
SKILL.md
技能文件保留在发布者的仓库中。
在终端中粘贴复制的命令。它会把这个 SKILL.md 安装到对应的智能体中。
npx skills add BrianRWagner/ai-marketing-claude-code-skills --skill voice-extractor -a cursor -y在终端中粘贴复制的命令。它会把这个 SKILL.md 安装到对应的智能体中。
npx skills add BrianRWagner/ai-marketing-claude-code-skills --skill voice-extractor -a claude-code -y在终端中粘贴复制的命令。它会把这个 SKILL.md 安装到对应的智能体中。
npx skills add BrianRWagner/ai-marketing-claude-code-skills --skill voice-extractor -a codex -y在终端中粘贴复制的命令。它会把这个 SKILL.md 安装到对应的智能体中。
npx skills add BrianRWagner/ai-marketing-claude-code-skills --skill voice-extractor -a grok -y在 Grok Bot 聊天中发送复制的消息。Bot 会从 GitHub 获取这个 SKILL.md,并添加到你所有 Bot 共享的技能中。
Install this skill: https://github.com/BrianRWagner/ai-marketing-claude-code-skills/tree/HEAD/voice-extractor来自发布者
AI-generated content all sounds the same. The fix isn't better prompts — it's teaching the AI how you actually communicate.
This skill extracts your communication DNA from writing samples and produces a Voice Guide: documented, tested, and ready to use.
Detect from context or ask: "Quick voice snapshot, full Voice Guide, or full guide with examples?"
Default: standard — use quick if they just need a fast reference. Use deep if they're onboarding a ghostwriter or building a content team.
Before extracting, collect:
Sample priority (most → least authentic):
Minimum sample gate: If samples total under 500 words, stop:
"These samples are too short to extract reliable patterns. Please add 2-3 more — emails, Slack messages, or transcripts work best. The messier and more casual, the better."
Do not attempt full extraction from under 500 words. Offer quick mode instead.
Before extracting, reason through:
Output a sample assessment:
"I have [X samples / Y words] to work with. Quality: [high/medium — why]. I'll use [full/quick] mode. Excluding: [any patterns and why]."
Identify the fundamental communication mode:
Role:
Default energy:
Recurring themes: What topics appear unprompted across samples? These are the things they actually care about.
Scan all samples and extract:
Transition phrases (how they shift topics):
Emphasis phrases (how they land a point):
Closers (how they wrap up):
Map their stated expertise areas to each zone. This calibration is what makes the voice feel real vs. one-dimensional.
Extract what they'd NEVER say:
Source these from sample evidence where possible: "You never used [word] across [X samples] — it doesn't fit your voice."
After extracting the full profile, generate 2 test sentences on the same topic:
Version A (using the extracted voice profile):
"[Sample sentence in their voice]"
Version B (wrong voice — contrasting example):
"[Same content, different voice — shows what to avoid]"
Ask the user: "Does Version A actually sound like you when you're not overthinking it? What feels off?"
This validation catches extraction errors before the guide is put into production.
--quick)When samples are thin (300–500 words) or time is short:
Output: Minimum viable voice guide.
Difference from full mode:
After generating the Voice Guide:
Flag any issues: "The anti-pattern section only has 2 entries — not enough for a usable guide. I need more samples or direct input from the user."
## Voice Guide: [Name] — [Date]
### Sample Assessment
- Samples: [count, types]
- Total words: [count]
- Quality: [high/medium — reason]
- Mode: [quick/full]
- Excluded: [patterns excluded + why]
---
### Core Energy
- Role: [teacher/challenger/cheerleader/straight-shooter]
- Default energy: [description]
- Recurring themes: [list]
### Signature Phrases
**Transitions:**
- "[Phrase]" (source: [email/post])
- "[Phrase]"
**Emphasis:**
- "[Phrase]" (source: [email/post])
**Closers:**
- "[Phrase]"
### Confidence Calibration
**Full authority (no hedging):**
Topics: [list]
Sounds like: "[example sentence]"
**Earned perspective:**
Topics: [list]
Sounds like: "[example sentence]"
**Active exploration:**
Topics: [list]
Sounds like: "[example sentence]"
### Anti-Patterns (Never Use)
- [Word/phrase] — why: [evidence from samples]
- [Word/phrase] — why: [evidence]
### Validation Test
**This sounds like you:**
"[Version A]"
**This doesn't:**
"[Version B — contrast]"
### Self-Critique Notes
[Any gaps, things to validate with user]
### Usage Instructions
- For AI: Paste this guide into your system prompt
- For ghostwriter: Share on day 1 — cuts revision cycles in half
- For team: This is the benchmark for "on brand"
Skill by Brian Wagner | AI Marketing Architect | brianrwagner.com
Extract and document someone's authentic writing voice from samples. Use when someone needs a "voice guide," wants to capture their writing DNA, or needs to train AI to write in their style. Also usef
---name: voice-extractordescription: Extract and document someone's authentic writing voice from samples. Use when someone needs a "voice guide,"…---# voice-extractor/SKILL.md概览
Extract and document someone's authentic writing voice from samples. Use when someone needs a "voice guide," wants to capture their writing DNA, or needs to train AI to write in their style. Also useful for ghostwriting, brand voice documentation, or onboarding writers.
常见问题
从发布者处下载 voice-extractor/SKILL.md,或复制路径后打开仓库。PlugMCP 只收录该技能,不托管文件。
发布者将它存放在 https://github.com/BrianRWagner/ai-marketing-claude-code-skills。
Claude、Cursor、Codex 和 Grok Build 可以读取你从发布者仓库添加的 SKILL.md。在聊天中发送技能的 GitHub 文件夹链接,Grok Bot 就会安装它。