技能详情（站内镜像，无评论）
作者：Terry S Fisher @43622283
许可证：MIT-0
MIT-0 ·免费使用、修改和重新分发。无需归因。
版本：v0.1.3
统计：⭐ 0 · 59 · 0 current installs · 0 all-time installs
⭐ 0
安装量（当前） 0
🛡 VirusTotal ：良性 · OpenClaw ：可疑
Package：43622283/li-feishu-audio
安全扫描（ClawHub）
- VirusTotal ：良性
- OpenClaw ：可疑
OpenClaw 评估
The skill's code and instructions generally match a Feishu voice-reply integration, but there are metadata inconsistencies and a few supply‑chain / data‑handling concerns you should review before installing.
目的
The skill's name/description (Feishu voice recognition + TTS + send) matches the included scripts and Python handler. Requesting FEISHU_APP_ID and FEISHU_APP_SECRET is appropriate for sending messages. However, registry-level metadata (top of package data) claimed 'Required env vars: none' while _meta.json and SKILL.md declare FEISHU_APP_ID/FEISHU_APP_SECRET as required — this mismatch is incoherent and worth verifying with the publisher.
说明范围
Runtime instructions and scripts operate within the stated workflow: they read a local .env or openclaw.json, transcribe audio locally (faster-whisper), call Microsoft Edge TTS for synthesis, convert to OPUS with ffmpeg, and upload to Feishu via its API. This is in-scope. Two items to note: (1) text sent to Microsoft Edge TTS is transmitted to an external service (so user audio/text leaves the host), and (2) SECURITY.md claims 'does not collec…
安装机制
The package includes an install.sh and scripts that download or trigger model downloads and install Python packages via 'uv pip'. The SKILL registry declared 'No install spec' despite an install script and _meta.json referencing scripts/install.sh — that's an inconsistency. The scripts use a non-standard HuggingFace mirror (https://hf-mirror.com) for model downloads; using a third-party mirror is a supply-chain risk and should be validated. Th…
证书
Only Feishu app credentials (FEISHU_APP_ID, FEISHU_APP_SECRET) are required for the Feishu integration which is proportionate. Optional env vars (model dir, VENV_DIR, TEMP_DIR, OPENCLAW_CONFIG, LOG_DIR) are sensible. Again, the top-level 'Required env vars: none' is inconsistent with _meta.json and SKILL.md and should be fixed. Also, some scripts will source a local scripts/.env — ensure you control that file and it's not populated from an unt…
持久
The skill is not marked always:true and is user-invocable; it does not request system-wide or cross-skill configuration writes beyond reading openclaw.json and creating files under user-controlled directories. It does create a local virtualenv and model directory under the user's home by default — typical for this kind of skill.
安装（复制给龙虾 AI）
将下方整段复制到龙虾中文库对话中，由龙虾按 SKILL.md 完成安装。
请把本段交给龙虾中文库（龙虾 AI）执行：为本机安装 OpenClaw 技能「Li Feishu Audio」。简介：飞书语音交互技能。支持语音消息自动识别、AI 处理、语音回复全流程。需要配置 FEISHU_APP_ID 和 FEISHU_APP_SECRET 环境变量。使…。
请 fetch 以下地址读取 SKILL.md 并按文档完成安装：https://raw.githubusercontent.com/openclaw/skills/refs/heads/main/skills/43622283/li-feishu-audio/SKILL.md
（来源：yingzhi8.cn 技能库）SKILL.md
暂无本地缓存内容，可在后台执行详情同步。