← 返回 Skills 市场
128
总下载
0
收藏
0
当前安装
2
版本数
在 OpenClaw 中安装
/install openclaw-voice-patch
功能描述
Patch OpenClaw Control UI to add dual voice input buttons (auto-send + continuous). Use when OpenClaw voice input needs patching, after OpenClaw updates (pat...
安全使用建议
This skill appears to do what it says (patch a local OpenClaw JS file), but follow these precautions before proceeding:
- Do not run sudo chown -R commands or copy files into the agent workspace without understanding exactly which path will be changed. Prefer to perform privileged actions yourself in a terminal rather than giving files to an agent.
- Make a backup (the skill recommends cp file file.bak). Keep multiple backups before editing.
- Inspect the index-<hash>.js file yourself (open in an editor) and review the exact replacement text the skill will apply. If you don't understand the JS, ask a developer you trust to review it.
- Instead of copying the file into the agent workspace, consider applying the patch locally (manual find/replace or a script you run) and then share only a diff or summary with the agent for verification.
- Run the patched app in a sandbox or test environment first to confirm behavior. If microphone permission policy changes are required, make those changes deliberately and minimally.
If you want a safer path, provide the exact index-<hash>.js file (only after removing unrelated sensitive content) to a trusted developer, or ask the skill author to provide a small, auditable patch script you run locally rather than giving the agent write access.
功能分析
Type: OpenClaw Skill
Name: openclaw-voice-patch
Version: 1.0.1
The skill instructs the AI agent to perform high-risk operations, including searching the entire filesystem ('find /' or 'Get-ChildItem -Path C:\ -Recurse') and modifying production JavaScript assets ('index-*.js') of the OpenClaw installation. It also includes instructions for the agent to prompt the user for elevated permissions ('sudo chown') or to move system files into the agent's workspace if access is denied. While these actions are aligned with the stated goal of patching the voice input UI, the broad file access and modification of application code in 'SKILL.md' constitute significant security risks.
能力评估
Purpose & Capability
Name/description match the requested actions: the skill only needs to find and modify a local OpenClaw UI JS asset. There are no unrelated environment variables, binaries, or external services requested.
Instruction Scope
The SKILL.md tells the agent to search the entire filesystem for index-<hash>.js and to edit that file in place. It also instructs the user to either copy the file into the agent workspace or to run sudo chown -R on the OpenClaw dist path. Asking the user to copy potentially sensitive local files into the agent workspace or to change file ownership are privacy/privilege risks and broaden the agent's access beyond purely patching code. The patching steps themselves are specific (text patterns to match/replace), which is good, but the guidance to grant write access or move files into the workspace is risky and should be handled carefully.
Install Mechanism
Instruction-only skill with no install spec and no code files; nothing is written to disk by an installer. This is the lowest-risk install model.
Credentials
No environment variables, credentials, or external endpoints are requested. The only 'permissions' asked for are local filesystem access to the OpenClaw install (copying files, backing up, or changing ownership), which are proportional to a file-patching task but still sensitive.
Persistence & Privilege
The skill is not flagged as always:true and does not request persistent platform privileges. However, the instructions explicitly suggest using sudo chown -R to grant write access — that is a user action that elevates privileges for files and should be avoided unless the user understands the implications and limits the scope precisely to the OpenClaw dist directory.
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 openclaw-voice-patch - 安装完成后，直接呼叫该 Skill 的名称或使用
/openclaw-voice-patch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
v1.0.1
- Voice input patch now uses a single mic button with smart dual-mode: short click for auto-send, or hold 3 seconds for continuous mode.
- Tooltip adapts to state: shows "Voice (hold 3s for continuous)" when idle, and "Stop voice auto"/"Stop voice continuous" when recording.
- Auto-send mode automatically stops after pause or 3 seconds of silence (acts as misclick safeguard).
- Continuous mode restarts recognition after each message until manually stopped.
- Live interim transcript appears above the input in both modes.
- Documentation updated to reflect new hold-to-activate workflow; previous dual-button UI removed.
v1.0.0
- Initial release of openclaw-voice-patch.
- Adds two voice input modes to OpenClaw Control UI: "Voice auto-send" (one-shot) and "Voice continuous" (keeps recording for multiple messages).
- Includes detailed patch instructions, verification steps, and backup/rollback guidance for compatibility with OpenClaw updates.
- Describes necessary troubleshooting steps and browser support requirements.
- Provides clear guidance if the patch cannot access or modify the necessary files.
元数据
常见问题
Voice Input Patch – Dual Mic Buttons 是什么？
Patch OpenClaw Control UI to add dual voice input buttons (auto-send + continuous). Use when OpenClaw voice input needs patching, after OpenClaw updates (pat... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 128 次。
如何安装 Voice Input Patch – Dual Mic Buttons？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install openclaw-voice-patch」即可一键安装，无需额外配置。
Voice Input Patch – Dual Mic Buttons 是免费的吗？
是的，Voice Input Patch – Dual Mic Buttons 完全免费，采用 MIT-0 许可证，可自由下载、安装和使用。
Voice Input Patch – Dual Mic Buttons 支持哪些平台？
Voice Input Patch – Dual Mic Buttons 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 Voice Input Patch – Dual Mic Buttons？
由 ErForscht（@elzershark）开发并维护，当前版本 v1.0.1。
推荐 Skills