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 agentmods add skills/iflytek/ifly-skills/iflytek-hyper-ttsnpx skills add iflytek/iFly-Skills --skill iflytek-hyper-ttsgit clone --depth 1 https://github.com/iflytek/iFly-SkillsWhat 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 | $0.00092 | $0.04266 |
| Opus 5 | $0.00046 | $0.02133 |
| Sonnet 5 | $0.00018 | $0.00853 |
| Haiku 4.5 | $0.00009 | $0.00427 |
Grade A, and why
iflytek-hyper-tts 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 yesterday.
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 — 307 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Ifly Hyper TTS (讯飞超拟人语音合成)
将文本转换为超拟人语音。适用于"把这段文案读出来"等语音合成需求。
API 文档：https://www.xfyun.cn/doc/spark/super%20smart-tts.html
核心特性
- 文本转语音：通过 WebSocket 双向流式接口合成音频
- 发音人控制：通过
--vcn指定已授权的超拟人发音人，默认使用聆小糖 - 基础韵律控制：语速（speed）/ 音量（volume）/ 语调（pitch）
- 多属性控制：语言/方言通过选择不同发音人(VCN)实现
- 多格式输出：MP3（lame）
使用场景（AI调用指南）
场景一：用户说"把这段文案读出来"
用户直接提供文本，要求语音合成。默认使用系统预设参数。
触发条件：用户指令包含"读出来"、"念出来"、"语音合成"、"转语音"、"文字转语音"等关键词。
默认参数：
- 发音人：聆小糖 (
x5_lingxiaotang_flow) - 女声，中文普通话，最适合作为默认主音色 - 语速：50（正常）
- 音量：50（正常）
- 语调：50（正常）
- 输出格式：MP3（lame）
- 采样率：24000Hz
# 直接合成（使用默认参数）
python3 scripts/xfei_hyper_tts.py --text "你好，欢迎使用讯飞超拟人语音合成！"
# 指定输出文件
python3 scripts/xfei_hyper_tts.py --text "欢迎收听" --output welcome.mp3
# 使用默认发音人聆小糖
python3 scripts/xfei_hyper_tts.py --text "你好" --vcn x5_lingxiaotang_flow
场景二：用户明确指定参数
用户明确指定发音人、语速、语调、音量等参数时，按用户指定参数执行。
触发条件：用户明确提到发音人名称/VCN、语速数值、语调/音量调整等。
# 指定发音人（需填写已授权的 vcn 代码，可通过 list_voices 查看）
python3 scripts/xfei_hyper_tts.py --text "你好" --vcn x5_lingfeiyi_flow
# 调整语速
python3 scripts/xfei_hyper_tts.py --text "语速稍快" --output fast.mp3 --speed 70
python3 scripts/xfei_hyper_tts.py --text "语速稍慢" --output slow.mp3 --speed 30
# 调整音量和语调
python3 scripts/xfei_hyper_tts.py --text "大声一点" --output loud.mp3 --volume 80
python3 scripts/xfei_hyper_tts.py --text "音调高低" --output pitch_test.mp3 --pitch 60
场景三：语言/方言控制
⚠️ 重要说明：讯飞超拟人API不支持 --language、--style、--emotion 参数。语言和方言通过**选择不同发音人(VCN)**来实现。
# 中文普通话（默认）
python3 scripts/xfei_hyper_tts.py --text "你好"
# 英语
python3 scripts/xfei_hyper_tts.py --text "Hello" --vcn x5_EnUs_Grant_flow
# 天津话
python3 scripts/xfei_hyper_tts.py --text "干嘛" --vcn x4_zijin_oral
# 东北话
python3 scripts/xfei_hyper_tts.py --text "嘎哈呢" --vcn x4_ziyang_oral
场景四：查看可用发音人列表
python3 scripts/xfei_hyper_tts.py --action list_voices
发音人说明
精选音色池（Skill 默认支持）
Skill 默认只开放以下 7 个精选发音人，避免全部暴露：
| 姓名 | VCN | 性别 | 语言 | 适用场景 |
|---|---|---|---|---|
| 聆小糖 | x5_lingxiaotang_flow | 女声 | 中文普通话 | 默认主音色，适合语音助手 |
| 聆飞瀚 | x6_lingfeihan_pro | 成年男 | 中文普通话 | 纪录片、偏正式表达场景 |
| 温暖磁性男声 | x6_wennuancixingnansheng_mini | 成年男 | 中文普通话 | 角色配音、客服场景 |
| Grant | x5_EnUs_Grant_flow | 男 | 英语美式 | 英文场景 |
| Lila | x5_EnUs_Lila_flow | 女 | 英语美式 | 英文场景 |
| 天津话 | x4_zijin_oral | 成年男 | 天津话 | 方言场景 |
| 东北话 | x4_ziyang_oral | 成年男 | 东北话 | 方言场景 |
What ships with it
4 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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.
- yesterday First seen · 307 lines · 92 tokens per session scan A c4dac768dcf9
iflytek-hyper-tts is a skill published in the GitHub repository iflytek/iFly-Skills (214 stars, last pushed 2d ago), licensed Apache-2.0. It adds 92 tokens to every session and 4,266 once invoked, about $0.0005 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
scammer
骗子.skill — 把骗局套路蒸馏成 AI，检测可疑消息，定位当前阶段，预判下一步。Distill scam patterns into AI, detect suspicious messages, locate current phase, predict next move.
rewrite-meeting-audio-transcription
Rewrite raw meeting audio transcriptions into clean, accurate meeting minutes in Traditional Chinese. Use when the user has an unprocessed audio transcription file with recognition errors and needs it cleaned up into proper meeting minutes.
design-taste-frontend-v1
The original v1 taste-skill, preserved for projects depending on its exact behavior. The current default is design-taste-frontend (v2 experimental), which is a substantial rewrite. Use this v1 install name only if you need exact backward compatibility.
brandkit
Premium brand-kit image generation skill for creating high-end brand-guidelines boards, logo systems, identity decks, and visual-world presentations. Trained for minimalist, cinematic, editorial, dark-tech, luxury, cultural, security, gaming, developer-tool, and consumer-app brand systems. Optimized for intentional…
high-end-visual-design
Teaches the AI to design like a high-end agency. Defines the exact fonts, spacing, shadows, card structures, and animations that make a website feel expensive. Blocks all the common defaults that make AI designs look cheap or generic.
file_reader
Read and summarize text-based file types only. Prefer readfile for text formats; use executeshellcommand for type detection when needed. PDF/Office/images/archives are handled by other skills.