首页 › 龙虾技能列表 › Aliyun Asr — 技能工具

🎙️ Aliyun Asr — 技能工具

v1.0.10

[自动翻译] Pure Aliyun ASR skill for voice message transcription, supports multiple channels including Feishu

2· 2,038·7 当前·10 累计
by @jixsonwang (Jixson)·MIT-0
下载技能包
License
MIT-0
最后更新
2026/4/11
安全扫描
VirusTotal
可疑
查看报告
OpenClaw
可疑
medium confidence
The skill appears to implement Aliyun ASR functionality, but there are inconsistencies between its description/instructions and the actual code (notably a required local config file path, an undeclared ffmpeg dependency, and misleading claims about "no extra configuration"), so proceed with caution.
评估建议
This skill's code implements Aliyun ASR and calls official Aliyun endpoints, but there are important inconsistencies you should consider before installing: - The README's "no extra configuration" claim is false: you must create /root/.openclaw/aliyun-asr-config.json containing your Aliyun access_key_id/access_key_secret and app_key. The registry metadata did not declare this config path. Confirm you are comfortable storing credentials on disk at that location and that the agent process has perm...
详细分析 ▾
ℹ 用途与能力
Name/description match the code: the Python code calls Aliyun NLS endpoints to convert audio to text. However, the metadata claimed no required config paths or credentials while the implementation requires a settings file at /root/.openclaw/aliyun-asr-config.json containing AccessKeyId/Secret and app_key. The use of ffmpeg for format conversion is present in code but not declared in required binaries. These gaps are inconsistent with the published metadata/README.
⚠ 指令范围
SKILL.md asserts "automatic integration, no additional configuration" and "no data storage," yet runtime instructions and code require creating a config file with credentials under /root/.openclaw and advise chmod 600. The code will read that file and exit if missing. The handler also invokes ffmpeg via subprocess to convert OGG→WAV, and posts raw audio bytes to Aliyun endpoints. The README's automatic/zero-config claim is therefore misleading and grants the skill implicit access to a sensitive on-disk config path.
ℹ 安装机制
No install spec (instruction-only installer) — lower risk because nothing is auto-downloaded. The package includes Python code and declares dependency on the requests Python package in the README. However, ffmpeg is invoked at runtime but not listed as a required binary. There is also an empty index.js/package.json present (benign but unnecessary).
ℹ 凭证需求
The skill does not request environment variables but requires permanent credentials stored in a local JSON config file (access_key_id and access_key_secret). Those credentials are appropriate for calling Aliyun ASR, but storing them in /root/.openclaw implies the skill expects root-level file access. The number/type of secrets (Aliyun keys) is proportionate to the stated purpose, but the mismatch between declared/actual config requirements and use of a root path is concerning.
✓ 持久化与权限
Skill does not request always:true and does not modify other skills or system-wide settings. It runs as an on-demand handler and prints or returns recognized text. No indications of privileged persistence beyond reading the expected config file.
安全有层次，运行前请审查代码。

License

MIT-0

可自由使用、修改和再分发，无需署名。

运行时依赖

无特殊依赖

版本

latestv1.0.102026/2/10

安全修复：移除潜在敏感文件，添加index.js，清理缓存目录，确保无密钥泄露风险

● 可疑

安装命令 点击复制

官方npx clawhub@latest install aliyun-asr
镜像加速npx clawhub@latest install aliyun-asr --registry https://cn.clawhub-mirror.com

技能文档

纯语音识别，无语音合成 - 这是一个专门为OpenClaw设计的轻量级阿里云语音识别技能，只做一件事：将语音消息转换为文本。

🎯 核心功能

  • ✅ 纯ASR识别: 只进行语音到文本的转换，不生成任何语音回复
  • ✅ 多通道支持: 支持飞书(Feishu)、Telegram、WhatsApp等所有OpenClaw支持的语音消息通道
  • ✅ 自动集成: 无需额外配置，语音消息自动被识别并作为文本消息处理

⚙️ 快速配置

1. 阿里云准备

  • 开通 智能语音交互(NLS) 服务
  • 在RAM控制台创建子用户并分配 AliyunNLSFullAccess 权限
  • 在NLS控制台创建应用，获取 AppKey

2. 配置文件

创建配置文件 /root/.openclaw/aliyun-asr-config.json:

{
  "access_key_id": "your-access-key-id",
  "access_key_secret": "your-access-key-secret",
  "app_key": "your-app-key",
  "region": "cn-shanghai"
}

3. 安全设置

chmod 600 /root/.openclaw/aliyun-asr-config.json

🚀 使用方法

自动模式（推荐）

  • 用户向任何支持的通道发送语音消息
  • OpenClaw自动调用此技能识别语音内容
  • 识别的文本作为用户消息传递给AI
  • AI生成纯文本回复（不是语音）

🔧 技术细节

  • 依赖: requests (Python包)
  • 支持格式: MP3, WAV, OGG, FLAC, AMR, OPUS
  • API区域: 默认 cn-shanghai（可配置）

🛡️ 安全与合规

  • 无数据存储: 语音数据不存储在本地
  • 最小权限: 使用RAM子账号，避免主账号密钥
  • 配置分离: 敏感信息与代码完全分离

💡 开发规范

此技能严格遵循以下开发准则：

  • ✅ 完全符合开源skills的配置要求
  • ✅ 完全符合当地的法律法规要求
  • ✅ 未开发或未实现的功能，不包含在源码中
  • ✅ 本地测试代码，测试用例不包含在源码中
  • ✅ 密钥/认证隐私信息，不包含在源代码中
数据来源：ClawHub ↗ · 中文优化：龙虾技能库
OpenClaw 技能定制 / 插件定制 / 私有工作流定制

免费技能或插件可能存在安全风险，如需更匹配、更安全的方案，建议联系付费定制

了解定制服务