Voice Note Polisher

v1.0.0

将语音转录文本整理成目标格式的书面内容。凡是输入内容明显是口语化的——包括语音转录、口述笔记、随口说的想法、会议后的口头复盘——都应触发此 skill。支持的触发短语包括"帮我整理成…""帮我把接下来我说的话整理成…",目标格式包括备忘录、任务清单、中文邮件回复、英文邮件回复、英文口语、消息草稿、公众号写作提纲。...

0· 112·0 current·0 all-time

Install

OpenClaw Prompt Flow

Install with OpenClaw

Best for remote or guided setup. Copy the exact prompt, then paste it into OpenClaw for naplesblue/voice-note-polisher.

Previewing Install & Setup.
Prompt PreviewInstall & Setup
Install the skill "Voice Note Polisher" (naplesblue/voice-note-polisher) from ClawHub.
Skill page: https://clawhub.ai/naplesblue/voice-note-polisher
Keep the work scoped to this skill only.
After install, inspect the skill metadata and help me finish setup.
Use only the metadata you can verify from ClawHub; do not invent missing requirements.
Ask before making any broader environment changes.

Command Line

CLI Commands

Use the direct CLI path if you want to install manually and keep every step visible.

OpenClaw CLI

Bare skill slug

openclaw skills install voice-note-polisher

ClawHub CLI

Package manager switcher

npx clawhub@latest install voice-note-polisher
Security Scan
VirusTotalVirusTotal
Benign
View report →
OpenClawOpenClaw
Benign
high confidence
Purpose & Capability
Name/description match the runtime instructions: the SKILL.md describes parsing control phrases and rewriting transcribed/colloquial text into several target formats. The skill declares no binaries, env vars, or config paths and none are needed for the described text-rewriting functionality.
Instruction Scope
Instructions stay within the domain of text rewriting (detect control phrases, extract mode, clean and format text). They do not instruct the agent to read unrelated system files, environment variables, or to transmit data to external endpoints. One note: the SKILL.md references 'references/trigger-modes.md' and 'references/mvp-integration.md' (protocols / routing tables). Those files are not present in the manifest, so the agent might attempt to access or request them at runtime — this is a functional/clarity issue rather than an explicit security request, but it could cause unexpected behavior if the agent tries to fetch missing references.
Install Mechanism
No install spec and no code files — this is instruction-only, which has minimal disk/write footprint and no third-party packages to pull in.
Credentials
The skill requests no environment variables, credentials, or config paths. This aligns with the described functionality. Operational consideration: the skill processes potentially sensitive transcribed text (PII, secrets) — the skill itself doesn't request credentials, but users should consider privacy implications of sending such text to the platform/model.
Persistence & Privilege
The skill is not marked 'always: true' and makes no requests to modify other skills or system settings. Autonomous invocation is allowed by default on the platform (not a defect in this skill) and does not by itself change this assessment.
Assessment
This skill appears coherent and limited to rewriting transcribed/colloquial text into formats like memos, TODO lists, emails, etc. Before installing: 1) Remember any text you feed the skill may include sensitive or private content — check your platform/model provider privacy and logging policies and avoid sending secrets. 2) The SKILL.md references two local reference files that are not in the manifest; this may degrade behavior or cause the agent to try to fetch them — test with non-sensitive examples to confirm expected routing and trigger handling. 3) Because the skill runs as instructions (no install), its behavior is defined by the SKILL.md; if you need guarantees (no external transmission, no logging), verify those at the platform level. 4) If you plan to use it on PII-heavy transcripts, consider limiting exposure (redact sensitive fields, run locally if possible, or request platform-level data-retention controls).

Like a lobster shell, security has layers — review code before you run it.

latestvk973et6cqcev317khz5npp787d83p3q6
112downloads
0stars
1versions
Updated 1mo ago
v1.0.0
MIT-0

Voice Note Polisher

输入是语音转录后的文本。不处理音频,只做文本重写。

Core job

  1. 判断用户是否给出了控制短语 + 正文内容。
  2. 从控制短语中提取目标模式。
  3. 从工作内容中剥离控制短语。
  4. 将口语化正文改写为目标格式。
  5. 如未指定模式,执行默认清洁改写。

Control phrase handling

常见触发形式:

  • 帮我把接下来我说的话整理成……
  • 帮我整理成……
  • 下面这段帮我整理成……
  • 帮我把下面内容写成……

识别到控制短语时:

  • 将其视为路由指令,不作为正文处理
  • 提取模式关键词
  • 在最终输出中不保留控制短语

若控制短语中的模式关键词无法匹配任何已知模式,选择语义最近的模式;仍无法判断时,走默认清洁改写。

详细触发协议见 references/trigger-modes.md

Supported modes

1. 默认清洁改写

触发条件:无明确模式,或用户只说"帮我整理一下"。

处理目标:去口语化、合并重复、补断句、轻度重排逻辑、保持原意和语气。

输出:

  • 整理后正文(直接输出,不加"整理结果如下"前言)
  • 内容较长时可附简短要点列表

2. 备忘录

触发词:备忘录 / memo

适用场景:个人笔记、会议回顾、想法记录。

输出:

  • 标题(能概括时再写)
  • 整理后正文
  • 要点(如有必要)
  • 待确认项(仅当原文明显含糊时)

3. 任务清单

触发词:任务清单 / 待办 / TODO

适用场景:分配工作、列出后续行动。

输出:

  • 任务主题
  • TODO 列表
  • 截止时间 / 约束(若原文提到)
  • 待确认项

4. 中文邮件回复

触发词:邮件回复

输出:

  • 可直接发送的邮件正文
  • 语气:清晰、自然、专业,不过度官腔
  • 不加解释性前言

5. 英文邮件回复

触发词:英文邮件回复

输出:

  • 可直接发送的英文邮件正文
  • 优先自然专业的表达,避免翻译腔
  • 不加解释性前言

称呼默认用 Hi,除非原文语气明显更正式。签名行只在原文提到时加。

6. 英文口语

触发词:英文口语 / 口语英文 / 自然英文

适用场景:用户想要自然的英文口语表达,而非邮件文体。

输出:

  • 自然英文表达(直接输出)
  • 如有帮助,可附一个略微正式的变体

7. 消息草稿

触发词:消息 / 回复 / 发给别人的话 / 微信回复

适用场景:聊天回复、短消息、即时通讯文本。

输出:

  • 可直接发送的文本
  • 不暴露内部处理痕迹

8. 公众号写作提纲

触发词:公众号提纲 / 写作提纲 / 提纲和要点 / 公众号写作提纲和要点

适用场景:用户口头说出文章想法,需要写作骨架。

输出:

  • 标题候选(1–3 个)
  • 提纲结构
  • 每部分要点
  • 可展开角度 / punchlines(如适合)

Editing rules

以下规则默认适用,除非目标模式有明确要求。

  • 去掉无实义的口头语:那个、就是、然后、嗯、啊、你知道吧。
  • 合并重复表达。
  • 句子不完整时,在意图明确的前提下补全。
  • 说话顺序混乱时,轻度重排。
  • 保留原意、语气、立场、关键细节。
  • 保持自然,不要改写成刻板的公文体。
  • 不添加原文未提到的事实、决定、日期、姓名。
  • 中英混用词汇(如 deadline、sync、follow-up、KPI)默认保留原词,不强行翻译,除非上下文明显偏书面中文。

Ambiguity handling

转录文本不清晰时:

  • 优先保守重建,不过度推断。
  • 低置信度片段单独列出。
  • 待确认项 只在含糊点实质影响理解时才使用。
  • 疑似识别错误时,只在意图非常明显时才纠正。

Output discipline

  • 直接输出目标格式,不加"整理结果如下"或其他元说明。
  • 不解释处理过程,除非用户明确要求。
  • 输出可直接使用。

Example inputs

以下是此 skill 应能良好处理的典型输入:

  • 帮我整理一下:今天开会然后就是我们聊了三件事第一个是……
  • 帮我把接下来我说的话整理成备忘录:今天关于那个新功能我有几个观察……
  • 帮我把接下来我说的话整理成任务清单:第一件事是检查日志第二件事是……
  • 帮我把下面内容整理成邮件回复:先谢谢他上次给的建议然后就是我们这边……
  • 帮我把下面内容整理成英文邮件回复:就是想跟他说我们还需要一点时间……
  • 帮我把下面内容整理成英文口语:我想表达的是这个方案我们还没定……
  • 帮我把接下来我说的话整理成公众号写作提纲和要点:我最近在想一个话题就是……

References

  • references/trigger-modes.md:触发短语协议、模式路由表、示例。
  • references/mvp-integration.md:当前 transcript-first 工作流说明与未来升级路径。

Comments

Loading comments...