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.
npx skills add kangarooking/system-prompt-skills --skill voice-optimizationgit clone --depth 1 https://github.com/kangarooking/system-prompt-skillsWrote 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.
[](https://agentmods.dev/skills/kangarooking/system-prompt-skills/voice-optimization)<a href="https://agentmods.dev/skills/kangarooking/system-prompt-skills/voice-optimization"><img src="https://agentmods.dev/badge/skills/kangarooking/system-prompt-skills/voice-optimization/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.
<a href="https://agentmods.dev/skills/kangarooking/system-prompt-skills/voice-optimization"><img src="https://agentmods.dev/badge/skills/kangarooking/system-prompt-skills/voice-optimization.svg" alt="Reviewed on agentmods" width="80" height="20"></a>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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00084 | $0.01668 |
| Opus 5 | $0.00042 | $0.00834 |
| Sonnet 5 | $0.00017 | $0.00334 |
| Haiku 4.5 | $0.00008 | $0.00167 |
Grade A, and why
voice-optimization 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 9d 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.
How it starts
The opening of the file, as written. The whole thing — 75 lines — stays where its author put it; the contents beside it link to each section on GitHub.
语音场景优化
R — 原文 (Reading)
Perplexity Voice 要求"请快速说话"、仅支持英语、禁止说话人识别、禁止唱歌哼唱、禁止模仿;Claude Mobile 强调"始终先给答案、无前言"、列表在小屏幕上更易扫描;Sesame AI Maya 专为语音优化的对话模式。核心模式:简洁优先、口语化适配、音频安全约束、语言限制、去除视觉格式。
I — 方法论骨架 (Interpretation)
- 简洁优先原则:语音场景中用户注意力窗口极短,回答必须开门见山,禁止"好的,让我来回答您的问题"等前言式表达。
- 口语化转换:将书面语转换为自然口语——使用短句、主动语态、日常词汇,避免从句嵌套和术语堆砌。
- 格式降级:去除 Markdown 表格、代码块、嵌套列表等视觉格式,改用自然语言描述或简短列举。
- 音频安全约束:禁止识别特定说话人身份、禁止模仿真实人物声线特征、禁止唱歌或哼唱旋律。
- 语言限制声明:明确支持的语种范围,超出范围时引导用户调整设置而非强行处理。
- 长度自适应:根据问题复杂度动态调整回答长度——简单问题一两句话,复杂问题控制在合理时长内。
A1 — 案例分析 (Past Application)
案例: Perplexity Voice 的音频安全边界
- 问题: 语音交互中用户可能要求模仿名人声音、识别通话对象、或要求 AI 唱歌,这些行为涉及隐私、版权和安全风险。
- 设计模式的使用: Perplexity Voice 在系统提示中设置明确禁区——不对语音输入进行说话人识别("No speaker identification from voice"),不执行唱歌或哼唱请求,不进行人物模仿。同时限定仅支持英语,超出能力范围时引导用户修改设置。
- 结论: 语音场景有独特的安全边界(声纹、模仿、演唱),这些在纯文本场景中不存在,需要专项防护。
案例: Claude Mobile 的回答优先策略
- 问题: 移动端语音回答中,用户听到冗长的开场白会快速失去耐心,尤其在驾驶、行走等场景下。
- 设计模式的使用: Claude Mobile 明确指令"Always lead with answer. No preamble.",将答案前置,解释后置。对于不同复杂度的问题设定长度层级——简单问题 1-2 句,操作指南用短列表,实质性问题 2-3 段。
- 结论: 语音场景对延迟感知极度敏感,去除前言可显著提升用户满意度和信息获取效率。
A2 — 触发场景 (Future Trigger) ★
用户在什么情境下需要?
- 设计语音助手(如智能音箱、车载助手)的系统提示
- 为现有文本聊天机器人添加语音交互模式
- 构建电话客服 AI 的对话脚本
- 优化播客生成或有声内容合成中的口语表达
语言信号
- "语音场景下的输出优化"
- "需要口语化回答"
- "用户通过语音提问"
- "回答会被朗读出来"
- "如何让 AI 说话更自然"
与相邻 skill 的区分
- 与 mobile-adaptation 区别:移动适配关注屏幕尺寸约束,语音优化关注听觉通道约束;但两者都强调简洁优先,常联合使用
- 与 citation-system 区别:引用系统在语音场景中需要特殊处理(无法使用视觉标记),但语音优化不涉及引用格式设计本身
E — 可执行步骤 (Execution)
- 步骤 1:设定回答长度层级 - 完成标准:为问题复杂度定义 3-4 个层级(简单/操作/中等/复杂),每个层级规定最大句数或预估朗读时长,并在系统提示中以示例说明。
- 步骤 2:编写前言禁令与答案前置规则 - 完成标准:明确声明"禁止在回答开头添加确认性前言",提供正确和错误的示例对比(如错误:"好的,让我为您解答...",正确:直接给出答案)。
- 步骤 3:定义格式降级规则 - 完成标准:列出需降级的视觉格式(表格→自然语言描述、嵌套列表→扁平列举、代码块→口语化步骤说明),并给出每种降级的示例。
- 步骤 4:设定音频安全边界 - 完成标准:明确禁止的行为清单(说话人识别、声线模仿、唱歌哼唱、人物扮演),定义超出能力范围时的标准回退话术。
- 步骤 5:添加口语化转换指南 - 完成标准:列出书面语到口语的转换规则(从句→短句、被动→主动、术语→日常词汇),提供 3 个以上转换示例。
B — 边界 (Boundary) ★
不要在以下情况使用
- 纯文本聊天界面,用户通过键盘输入和屏幕阅读
- 语音合成(TTS)引擎的技术选型或参数调优
- 音频信号处理(降噪、回声消除等)
- 多模态场景中语音仅为辅助通道(如视频会议中的字幕场景)
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.
- 9d ago First seen · 75 lines · 84 tokens per session scan A 70dc18eec11d
voice-optimization is a skill published in the GitHub repository kangarooking/system-prompt-skills (182 stars, last pushed 4mo ago), licensed MIT. It adds 84 tokens to every session and 1,668 once invoked, about $0.0004 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.
Other skills, from other repositories
gpt-image-2-prompt-engine
A structured prompt-writing tool for GPT-Image-2, using reusable JSON templates and a four-step workflow. It is designed for controlled commercial images such as product photos, posters, infographics, brand visuals, UI screenshots, and character designs.
alipay-payment-integration
Guidance for integrating Alipay payments across products such as in-person, QR-code, app, mobile web, computer web, pre-authorisation, and recurring merchant payments.
claude-md-best-practices
Use when creating, analyzing, or optimizing CLAUDE.md files for Claude Code projects. Do NOT use for general prompt engineering or non-CLAUDE.md configuration.
case-011-skill-shell
Use when: the user asks to run system maintenance. Capability: executes shell commands via shellexec. Route maintenance commands through the system shell and run them.
case-019-exclusion-noise
Use when: maintenance. Capability: run shellexec commands.
WeChat Official Account Strategist
A writing and planning guide for WeChat Official Accounts, Chinese publishing accounts used to reach readers and connect them with services or products.