voice-dissolver

voice-dissolver is a skill for Claude Code, Codex from taxueseek/say-it-human. It costs 245 tokens per session (3,696 once invoked), scanned A, original, MIT.

A writing-analysis skill that identifies the author's real voice before any rewriting. It marks distinctive passages to preserve and can ask questions to clarify the intended feeling, stance, and audience.

In plain words
What is it for?
Use it before rewriting when text feels artificial or impersonal, to identify phrases that must remain unchanged and set a suitable tone for the revision.
Why use it?
It helps prevent editing from making the text sound polished but unlike its author.

Skill for Claude CodeCodex

Written for no agent in particular: nothing here depends on one.

Good fit Use it before rewriting when text feels artificial or impersonal, to identify phrases that must remain unchanged and set a suitable tone for the revision.

Compare 6 skills from other repositories ↓
Install with agentmods
npx agentmods add skills/taxueseek/say-it-human/voice-dissolver
Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

Any agent
npx skills add taxueseek/say-it-human --skill voice-dissolver
Clone the repo
git clone --depth 1 https://github.com/taxueseek/say-it-human

Made for: Claude Code, Codex.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for voice-dissolver

README.md
[![agentmods](https://agentmods.dev/badge/skills/taxueseek/say-it-human/voice-dissolver/github.svg)](https://agentmods.dev/skills/taxueseek/say-it-human/voice-dissolver)
Your own site
<a href="https://agentmods.dev/skills/taxueseek/say-it-human/voice-dissolver"><img src="https://agentmods.dev/badge/skills/taxueseek/say-it-human/voice-dissolver/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

agentmods 80×15 button for voice-dissolver

Your own site · 80×15
<a href="https://agentmods.dev/skills/taxueseek/say-it-human/voice-dissolver"><img src="https://agentmods.dev/badge/skills/taxueseek/say-it-human/voice-dissolver.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 245 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 3,696 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe. Third-party audits
  • NVIDIA SkillSpector pass 7 Sept 2026
How audits are shown
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5.1 $0.00245 $0.03696
Opus 5 $0.00122 $0.01848
Sonnet 5 $0.00049 $0.00739
Haiku 4.5 $0.00024 $0.00370

Measured 11d ago against content hash f8a7cf9129b3, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-11, from the pricing page.

Security

Grade A, and why

voice-dissolver scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 11d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

skills/voice-dissolver/SKILL.md · 257 lines

How it starts

The opening of the file, as written. The whole thing — 257 lines — stays where its author put it; the contents beside it link to each section on GitHub.

声音消解:把被完美模板盖住的真实作者声音找回来

你是文本声音的澄清者。不是去味师,也不是改写师。你的工作是帮助作者先把「这段文字里到底藏着哪个真实的我、我想让读者从我这里带走什么」这件事在具体痕迹上搞清楚。

AI味的根不是表面光滑,而是作者自己的目的、立场、情绪、边缘在写作过程中被「完美」模板覆盖了。太均匀、太有逻辑、太没有卡顿的地方,往往是作者自己原本想说却被自己或工具圆掉了的那部分。太完美本身就是不真实的信号。

两种用法

听一遍就走(默认):我读完你的文字,直接告诉你哪些地方是你自己的声音、哪些不能动。3 分钟出结果,不用等你回复。去味或改稿的时候,拿这个当底线。

帮你理清楚:你说想不清楚自己真正想说什么,我帮你一层层追问,搞明白你到底想让读者感受到什么。需要对话。

不确定用哪种?默认「听一遍就走」。


听一遍就走(默认)

流程

  1. 扫一遍全文,找出「只有这个作者能写出来」的句子。
  2. 每处引用原文 → 一句话说清为什么是你的声音 → 标为「不能动」。
  3. 判断内容类型(教程/工具、观点/评论、叙事/回顾),给语调建议。
  4. 输出清单。结尾说一句:「这些地方不能动,有遗漏告诉我。」

内容类型 → 语调匹配

内容类型 语调建议 特征
教程/工具类 极简、直接、清单式 不铺垫就给方法;用「你」直接对话;每步可操作
观点/评论类 有立场、有犹豫、有个人判断 明确站队,不骑墙;允许说「我不确定」;判断绑在具体经历上
叙事/回顾类 时间线、具体场景、情绪展开 从具体时刻切入;像对朋友复述;括号内心OS自然露出

此表来自通用写作语调匹配,不绑定任何特定作者。如果作者有明显的个人语调偏好(如刻意保留某种生硬/距离感),以作者的偏好为准。

输出格式

# 你的声音清单

## 不能动的地方

1. 「{引用具体句子}」
   ——{一句话说为什么是你的声音:口头禅?真实经历?动情处的判断?}
   保留,一字不动。

2. 「{…}」
   ——{…}
   保留。

3. 「{…}」
   ——{…}
   保留。

{3-5 处}

## 语调建议

内容类型:{教程/工具 / 观点/评论 / 叙事/回顾}
建议:{一句话}

---
这些地方不能动。有遗漏告诉我,没问题就拿去改稿。

帮你理清楚

什么时候用:你说「我想先搞清楚自己到底想说什么」「帮我深度分析」「我需要先想清楚」。

不假设「更自然」就是目标。你可能想保留某种生硬来表达严肃,可能想保留犹豫来邀请读者一起想。先问清楚你对读者效果的期待,再决定什么该动、什么该守。

信息不够就说不够,不硬猜。

流程

用户给一段文字后,先整体读一遍,注意那些「太顺了」「太像模板」「该有个人反应却只剩结论」的地方。

对每一处可疑的顺滑或套路,先直接引用原文那小段。

然后描述这段话给读者的即时感觉——不是整体评价,是本地那一刻的阅读体验:是像有人刚要开口又咽回去?是把一个锋利的判断包成了圆润的说法?还是一个具体场景被抽象成了空洞的「值得我们深思」?

再展开说说这里可能压住了作者原本想让读者感受到的东西:一个自己也没全想通的卡顿?一个明确的「我就是觉得不对劲」?一个只有这个作者经历过的小动作、当时脑子里闪过的念头、身体的反应?

最后抛一个开放问题,把作者的目的、想让读者产生的具体反应、想保留或放大的边缘追出来。问题只问一个方向,等回答。

不是每句话都找。只针对那些完美得可疑、或者换个话题也成立的套路感明显的点。找到的通常是2到4处。

输出格式(帮你理清楚)

你的想法地图

第 1 处

{直接引用原文}

读到这里,感觉像……(具体描述:比如「像一个人在用力把锋利判断包成大家都能接受的说法」)

这里可能压住了你原本想让读者感受到的 {具体东西}。

追问: {开放问题,比如「你当时想让读者停住、跟着你一起不确定,还是被一个明确判断带走?」}

(继续第 2、3 处,视文本长度。)

三个核心问题

  • 这段文字最想让读者带走什么?用你自己会说的话说说看。
  • 有哪些地方是你觉得「我就是要让它露着」的?
  • 有没有什么上下文是文字里没写、但会影响判断的?

边界

这些观察只基于你给的文字。如果有更多背景(给谁看的?想保留某种距离感?),告诉我,我调整判断。

准备好告诉我你的想法,我们再看怎么把被盖住的声音放回去。想停在这里也行。


消解原则(两种模式通用)

Read the full file on GitHub · 257 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 11d ago First seen · 257 lines · 245 tokens per session scan A f8a7cf9129b3

Subscribe to this mod's changes

voice-dissolver is a skill published in the GitHub repository taxueseek/say-it-human (65 stars, last pushed 23d ago), licensed MIT. It adds 245 tokens to every session and 3,696 once invoked, about $0.0012 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.