← 返回 Skills 市场
238
总下载
0
收藏
0
当前安装
1
版本数
在 OpenClaw 中安装
/install shumianyu
功能描述
将分段 asr 文本流在结束时统一整理成流畅书面语，并在用户指定 target_language 时输出译文。适用于语音听写、会议转写、采访整理、播客或视频字幕后处理等场景，尤其适合不需要实时字幕、只需要在整段结束后得到最终书面结果的任务。保持接口抽象、可替换，不绑定任何特定厂商 api。
安全使用建议
This skill is instruction-only and coherent: it defines an abstract interface to accumulate ASR chunks, produce a single cleaned written output, and optionally translate when a target language is given. It does not itself install code or request secrets. Before deploying in a real product, note: the SKILL.md intentionally avoids binding to any vendor — the concrete implementation you choose (speech/translation provider, cloud storage, logging) will determine where data is sent and what credentials are needed. Review any implementation for (1) which external services handle raw transcripts (privacy risk), (2) what credentials those services require, and (3) retention/logging policies for sensitive speech content. If you plan to wire this skill to third-party translators or ASR services, inspect those integrations separately for network endpoints, auth scopes, and data retention.
功能分析
Type: OpenClaw Skill
Name: shumianyu
Version: 1.0.0
The skill bundle is a text-processing utility designed to rewrite ASR (Automatic Speech Recognition) transcripts into formal written language and provide optional translation. The instructions in SKILL.md and references/interface-guide-zh.md focus entirely on buffering text chunks and performing linguistic cleanup once a session ends, with no evidence of data exfiltration, malicious execution, or prompt injection.
能力评估
Purpose & Capability
Name/description describe buffering ASR chunks and producing a single post-processed written output (and optional translation). The skill is instruction-only and requests no binaries, env vars, or other resources — all consistent with a vendor-agnostic interface/spec.
Instruction Scope
SKILL.md stays within scope: it describes receiving ASR chunks, buffering, detecting explicit end signals, rewriting to written form, and optional translation. It does not instruct reading unrelated files, accessing system credentials, or sending data to hidden endpoints.
Install Mechanism
No install spec and no code files are included; this is the lowest-risk form (instruction-only). Nothing will be written to disk or downloaded by the skill as provided.
Credentials
No environment variables, credentials, or config paths are required. The declared primary credential is none — proportional to the described abstract, vendor-agnostic role.
Persistence & Privilege
Skill does not request always:true or other elevated persistence. It does not instruct modifying other skills or system-wide settings.
如何使用
- 确保已安装 OpenClaw（本地或 Docker 部署）
- 在对话框中输入安装命令：
/install shumianyu - 安装完成后，直接呼叫该 Skill 的名称或使用
/shumianyu触发 - 根据 Skill 的参数说明提供必要输入，即可获得结构化输出
版本历史
v1.0.0
- Initial release of the asr-written-rewriter-stream skill.
- Accepts segmented ASR text streams and outputs a single, fluent written result only after receiving an explicit end signal.
- Optionally translates the finalized written result when a target_language is provided.
- Designed for post-processing meeting transcripts, interviews, podcasts, and similar scenarios—optimized for whole-segment finalization over real-time subtitle needs.
- Keeps interfaces abstract and compatible with any upstream ASR provider; not tied to any specific API or service.
- Respects and applies domain lexicons for consistent handling of names, technical terms, and abbreviations.
元数据
常见问题
shumianyu 是什么？
将分段 asr 文本流在结束时统一整理成流畅书面语，并在用户指定 target_language 时输出译文。适用于语音听写、会议转写、采访整理、播客或视频字幕后处理等场景，尤其适合不需要实时字幕、只需要在整段结束后得到最终书面结果的任务。保持接口抽象、可替换，不绑定任何特定厂商 api。 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 238 次。
如何安装 shumianyu？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install shumianyu」即可一键安装，无需额外配置。
shumianyu 是免费的吗？
是的，shumianyu 完全免费，采用 MIT-0 许可证，可自由下载、安装和使用。
shumianyu 支持哪些平台？
shumianyu 跨平台运行，可在任意部署了 OpenClaw / Claude Code 的环境中使用（cross-platform）。
谁开发了 shumianyu？
由 HeiMaoM（@hei-maom）开发并维护，当前版本 v1.0.0。
推荐 Skills