← 返回 Skills 市场

飞书语音回复

Name: 飞书语音回复
Author: jeffli2002

作者 jeffli2002 · GitHub ↗ · v1.0.0 · MIT-0

cross-platform ⚠ suspicious

137

总下载

当前安装

版本数

在 OpenClaw 中安装

/install jeffli-feishu-voice

功能描述

Generate Feishu-native voice replies with a playable pause/resume bar by synthesizing text, converting it with ffmpeg to Ogg/Opus, and sending it as a voice...

使用说明 (SKILL.md)

Feishu Voice Reply

Create Feishu replies that contain both normal text and a real voice bubble/player. Avoid sending raw mp3 files when the goal is in-chat playback.

Workflow

Draft the text reply first.
Keep the visible text and the spoken text aligned.
Use scripts/build_feishu_voice.py to synthesize Chinese speech with Edge TTS and convert it to Ogg/Opus through ffmpeg.
Send the text reply normally.
Send the generated .ogg file with the messaging tool as a voice message:
- mimeType: audio/ogg
- asVoice: true
If a message id is available, reply to the triggering message for both the text and the voice.

Defaults

Default language: Chinese
Default voice: zh-CN-XiaoxiaoNeural
Default channel behavior: Feishu text message + Feishu voice message

Rules

语音文件必须存放在 Workspace 下：使用 /root/.openclaw/workspace/temp/voice/ 目录，不要用 /tmp。否则飞书只显示文件路径而不是语音播放条。 \x3C\x3C\x3C\x3C\x3C\x3C\x3C HEAD =======
Emoji 自动过滤：脚本会自动过滤 emoji 字符再合成 TTS，显示的文本保留 emoji，但语音不会朗读 emoji（避免杂音）。无需手动处理。

8d2abf78b8490403831aae82052e8e107054b856

Always show the text reply as well; do not send voice-only unless the user explicitly asks for that.
Prefer concise spoken text. If the full answer is long, speak a compact summary and keep the full text visible.
Use this skill specifically for Feishu playback UX. If the user only wants downloadable audio, normal attachments are fine.
If synthesis succeeds but voice sending fails, tell the user clearly that Feishu voice delivery failed instead of pretending it worked.

Script

Build a Feishu-compatible voice file

Run:

python3 scripts/build_feishu_voice.py \
  --text "今天上海多云，气温十一度。" \
  --out-dir /root/.openclaw/workspace/temp/voice

The script prints JSON including:

ogg_path: send this to Feishu as the voice file (必须在 Workspace 下)
mp3_path: intermediate file
voice: selected Edge TTS voice

Sending pattern

After building the file:

Send the visible text reply.
Send ogg_path with:
- channel feishu
- path = generated ogg_path
- mimeType = audio/ogg
- asVoice = true

Resource

scripts/build_feishu_voice.py: deterministic synthesis + ffmpeg conversion for Feishu voice playback.

安全使用建议

This skill appears to do what it says (synthesize speech with edge-tts and convert via ffmpeg), but there are several red flags you should address before installing or using it: 1) SKILL.md contains unresolved git merge conflict markers — ask the author to clean and re-publish; 2) SKILL.md claims the script auto-filters emoji, but the provided script contains no emoji filtering logic — request clarification or a corrected script if you need that behavior; 3) The script requires the 'edge-tts' and 'ffmpeg' binaries (and optionally 'ffprobe') but the skill metadata doesn't declare these prerequisites — ensure these tools are installed and trusted on the host; 4) The README instructs storing files under /root/.openclaw/workspace/temp/voice — confirm that writing to that workspace is acceptable in your environment and that no sensitive files could be exposed; 5) Because this is instruction-only with an included script, inspect and run the script in a safe, isolated test environment first. If the author fixes the merge conflict, documents dependencies explicitly, and either implements the claimed emoji filtering or removes that claim, the skill becomes much more coherent. If you need help vetting a corrected version, provide the updated files.

功能分析

Type: OpenClaw Skill Name: jeffli-feishu-voice Version: 1.0.0 The skill is a legitimate utility for generating Feishu-compatible voice messages from text. The core logic in `scripts/build_feishu_voice.py` uses standard tools (edge-tts and ffmpeg) and follows secure coding practices, such as using argument lists in subprocess calls to prevent shell injection and sanitizing filenames. No indicators of data exfiltration, malicious execution, or harmful prompt injection were found.

能力评估

⚠ Purpose & Capability

The description and script align on using Edge TTS + ffmpeg to produce Ogg/Opus for Feishu playback, which is coherent. However the registry metadata declares no required binaries or config paths, while the script explicitly requires the 'edge-tts' binary and 'ffmpeg' (and optionally 'ffprobe'), and the README instructs using a specific workspace directory (/root/.openclaw/workspace/temp/voice) that isn't declared — this mismatch between claimed requirements and actual needs is a coherence issue.

⚠ Instruction Scope

SKILL.md gives a focused workflow for building and sending Feishu voice messages, but it contains unresolved git merge conflict markers (<<<<<<< HEAD and >>>>>>>) and claims 'Emoji 自动过滤：脚本会自动过滤 emoji' while the included Python script contains no emoji-filtering logic. The README also mandates storing files in a specific workspace path; the script will accept an out-dir but the documentation's hard requirement is not enforced or declared. These discrepancies mean the runtime instructions and implementation are not fully consistent.

ℹ Install Mechanism

There is no install spec (instruction-only), which is lower risk. But the shipped script depends on external binaries (edge-tts, ffmpeg, optional ffprobe) that are not installed or declared by the skill; the environment must provide them. This is an operational gap rather than an active install risk, but the skill should declare these prerequisites.

✓ Credentials

The skill requests no credentials or environment variables and the script does not access secrets or external endpoints; it only invokes local binaries and writes files to disk. That is proportional to the stated purpose.

✓ Persistence & Privilege

The skill is not always-enabled and does not request elevated or persistent privileges or modify other skills. It writes output files to a workspace directory (as documented), which is normal for this task.

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 jeffli-feishu-voice
安装完成后，直接呼叫该 Skill 的名称或使用 /jeffli-feishu-voice 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

首次发布：飞书语音回复Skill，支持Edge TTS生成语音并以原生播放条形式发送

元数据

Slug jeffli-feishu-voice

版本 1.0.0

许可证 MIT-0

累计安装 0

当前安装数 0

历史版本数 1

常见问题