← Back to Skills Marketplace
Elevenlabs
by
Oliver Drobnik
· GitHub ↗
· v1.3.4
5419
Downloads
2
Stars
28
Active Installs
12
Versions
Install in OpenClaw
/install elevenlabs
Description
Text-to-speech, sound effects, music generation, voice management, and quota checks via the ElevenLabs API. Use when generating audio with ElevenLabs or mana...
Usage Guidance
Install only if you are comfortable using your ElevenLabs API key from this environment. Store the key securely, avoid committing .env files, and only use voice-cloning samples that you are authorized to upload and clone.
Capability Analysis
Type: OpenClaw Skill
Name: elevenlabs
Version: 1.3.4
The OpenClaw AgentSkills bundle for ElevenLabs is classified as benign. The skill demonstrates strong security practices, notably the `_pathguard.py` script which strictly limits output file writes to the workspace or `/tmp` directories, and the `voiceclone.py` script which restricts input audio file reads to a designated sample directory. All network requests are directed exclusively to the legitimate ElevenLabs API (api.elevenlabs.io). Command execution via `subprocess.run` (e.g., for `ffmpeg` and `afplay`) correctly uses a list of arguments, preventing shell injection. No evidence of prompt injection, data exfiltration, persistence mechanisms, or obfuscation was found in any of the files, including `SKILL.md` and `SETUP.md`.
Capability Assessment
Purpose & Capability
The scripts match the stated purpose: text-to-speech, sound effects, music, dialogue, voice listing, quota checks, and voice cloning through ElevenLabs API endpoints.
Instruction Scope
The artifacts provide command examples and setup guidance; they do not include hidden instructions, prompt overrides, automatic execution, or deception.
Install Mechanism
Setup asks the user to install ffmpeg and the Python requests package manually; there is no bundled installer or automatic dependency execution.
Credentials
Network use is limited to ElevenLabs API calls, output writes are guarded to workspace or temporary paths, and voice-clone input reads are restricted to a sample directory unless the user explicitly supplies another directory.
Persistence & Privilege
The skill requires an ElevenLabs API key and setup describes persistent storage in shell profiles or .env/state locations, which is expected for this integration but sensitive.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install elevenlabs - After installation, invoke the skill by name or use
/elevenlabs - Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.3.4
fix: use /Users/oliver/clawd for workspace root to preserve symlink paths
v1.3.3
- Improved quota and usage reporting in `quota.py` for clearer output.
- Updates to `SETUP.md` for better clarity on prerequisites and setup.
- Minor documentation improvements in `SKILL.md` and code path checking script.
- Version bump to 1.3.3.
v1.3.2
Declare missing dependencies, env vars, and sensitivity in metadata
v1.3.1
Rename .clawdhubignore to .clawhubignore
v1.3.0
Refactor: move setup/prerequisites to SETUP.md, keep SKILL.md lean
v1.2.2
Remove .env loading from quota.py — use env vars only
v1.2.1
Security: output path sandboxing — all scripts (speech, sfx, music, dialogs) now validate -o paths stay within workspace or /tmp via shared _pathguard module.
v1.2.0
Security: removed --unsafe-allow-any-path flag from voiceclone.py. File reads are now always restricted to the sample directory.
v1.1.3
Harden quota dotenv loading (no workspace-wide .env)
v1.1.2
Security: voiceclone.py now only reads samples from a dedicated directory by default; requires explicit flag to allow arbitrary paths.
v1.1.1
Avoid reading home-directory .env; declare ffmpeg/afplay usage in metadata/docs
v1.1.0
Update metadata to openclaw format, add emoji, refresh README
Metadata
Frequently Asked Questions
What is Elevenlabs?
Text-to-speech, sound effects, music generation, voice management, and quota checks via the ElevenLabs API. Use when generating audio with ElevenLabs or mana... It is an AI Agent Skill for Claude Code / OpenClaw, with 5419 downloads so far.
How do I install Elevenlabs?
Run "/install elevenlabs" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is Elevenlabs free?
Yes, Elevenlabs is completely free (open-source). You can download, install and use it at no cost.
Which platforms does Elevenlabs support?
Elevenlabs is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created Elevenlabs?
It is built and maintained by Oliver Drobnik (@odrobnik); the current version is v1.3.4.
More Skills