Skill flagged — suspicious patterns detected
ClawHub Security flagged this skill as suspicious. Review the scan results before using.
Audio Cog
v1.0.10AI audio generation and text-to-speech powered by CellCog. Three voice providers (OpenAI, ElevenLabs, MiniMax), voice cloning, avatar voices, sound effects g...
⭐ 4· 4.8k·35 current·37 all-time
byCellCog@nitishgargiitd
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Suspicious
medium confidencePurpose & Capability
The SKILL.md describes an SDK-backed integration with CellCog and three upstream providers (OpenAI, ElevenLabs, MiniMax) and features like voice cloning and avatar uploads that normally require service credentials and explicit installation. However, the registry metadata declares no required environment variables, no primary credential, and no install steps. That omission is inconsistent with the claimed capabilities and raises questions about where/how authentication and provider keys are handled.
Instruction Scope
The instructions are mostly usage examples (Python snippets) and high-level guidance for generating audio; they do not themselves instruct reading arbitrary user files or exfiltrating data. However, the SKILL.md explicitly tells the agent to "read the cellcog skill for the full SDK reference — file handling, chat modes, timeouts, and more," delegating file- and credential-handling behavior to another skill that is not declared as a dependency. That delegation is vague and grants broad discretion to the runtime if the referenced skill is present.
Install Mechanism
This is instruction-only (no install spec, no code files). That lowers install-time risk because nothing is downloaded or written by this skill itself. The SKILL.md lists a dependency (cellcog) but does not provide an install step — the environment must already provide the SDK/package.
Credentials
Given the described integrations, one would expect at minimum a CellCog API key and possibly upstream provider keys (OpenAI, ElevenLabs, etc.) or a clear statement that CellCog proxies/authenticates on the user's behalf. The skill declares no required env vars or credentials. This mismatch is disproportionate and ambiguous: installing or running the skill may implicitly rely on credentials stored elsewhere (another skill, global agent config) without making that explicit.
Persistence & Privilege
The skill does not request always: true, does not declare writing to agent/system config, and is user-invocable with normal autonomous invocation allowed. No elevated persistence or system-wide modifications are declared.
What to consider before installing
Before installing or using this skill: 1) Ask the publisher how authentication is handled — does it require a CellCog API key or provider keys (OpenAI, ElevenLabs, MiniMax)? Where should those secrets be placed? 2) Inspect the separate 'cellcog' skill/documentation the SKILL.md references to see precise file-handling, credential, and network behavior; this skill delegates those responsibilities and does not declare them here. 3) Be cautious about voice cloning: do not upload voice samples you do not own, and confirm the privacy/retention policy for uploaded audio. 4) If you need to provide API keys, scope them narrowly, rotate them regularly, and avoid supplying unrelated credentials. If the publisher cannot clearly explain where credentials live and what the SDK will access, treat the skill as risky and do not install it in sensitive environments.Like a lobster shell, security has layers — review code before you run it.
latestvk978qedfnpvh66tdwq6c8njns184rfn6
License
MIT-0
Free to use, modify, and redistribute. No attribution required.
Runtime requirements
🎵 Clawdis
OSmacOS · Linux · Windows