Back to skill
Skillv1.0.0
ClawScan security
Ai Voiceover For Videos · ClawHub's context-aware review of the artifact, metadata, and declared behavior.
Scanner verdict
SuspiciousApr 21, 2026, 3:54 PM
- Verdict
- suspicious
- Confidence
- medium
- Model
- gpt-5-mini
- Summary
- The skill's declared purpose (AI voiceover for videos) mostly matches its behavior, but there are inconsistencies (metadata vs registry) and privacy-sensitive actions (uploading media to an external domain and reading a user config path) that you should understand before installing.
- Guidance
- This skill appears to do what it says (upload videos to a cloud render service and return a voiced-over video), but it contacts an external API (mega-api-prod.nemovideo.ai) and may read a user config folder (~/.config/nemovideo/). Before installing or using it: 1) Do not upload sensitive or private videos to this service unless you trust the domain and privacy policy. 2) Ask the skill author for source code or a homepage and clarification about the configPath requirement (why does it need ~/.config/nemovideo/?). 3) If you test it, use a throwaway NEMO_TOKEN or anonymous token and avoid running it in environments containing other credentials. 4) Prefer skills with a verifiable source or published package; absence of a homepage/source increases risk. If the author clarifies the configPath mismatch and provides source code or an official homepage, reassess — that could reduce the concern.
Review Dimensions
- Purpose & Capability
- noteName/description match the runtime instructions: the skill uploads videos and requests a cloud render pipeline. Requesting a single API token (NEMO_TOKEN) is coherent for a cloud service. However, the SKILL.md frontmatter also declares a required config path (~/.config/nemovideo/) while the registry metadata listed no required config paths — this mismatch is unexplained and worth clarifying.
- Instruction Scope
- concernRuntime instructions direct the agent to: look for NEMO_TOKEN in the environment, create an anonymous token via POST if missing, upload user video files to https://mega-api-prod.nemovideo.ai, stream/poll SSE endpoints, and detect the agent install path to set attribution headers. Uploading user files to an external service is expected for this task but is privacy-sensitive. Detecting install paths and accessing a user config directory (~/.config/nemovideo/) are extra filesystem probes that are not necessary for simple voiceover generation and expand the skill's access to local data.
- Install Mechanism
- okNo install spec and no code files (instruction-only). This minimizes on-disk persistence and reduces code-install risk.
- Credentials
- concernThe skill only declares NEMO_TOKEN as the primary credential, which is proportionate. But the SKILL.md instructs creating or retrieving a token automatically if none exists and also references reading a user config path (~/.config/nemovideo/). That combination could allow the skill to read existing local credentials or config and transmit them to the external service — the registry's omission of the configPath makes this behavior unclear and concerning.
- Persistence & Privilege
- okalways:false and no install steps. The skill doesn't request permanent platform-level presence or modification of other skills. Autonomous invocation is allowed (platform default) but not, by itself, an elevated privilege.