Skillv1.0.0

ClawScan security

Video Generator Free From Text · ClawHub's context-aware review of the artifact, metadata, and declared behavior.

Scanner verdict

SuspiciousApr 13, 2026, 4:27 PM

Verdict: suspicious
Confidence: medium
Model: gpt-5-mini
Summary: The skill's described purpose (turn text into videos via a cloud backend) aligns with its runtime instructions, but there are small yet meaningful inconsistencies in the metadata and token handling that you should understand before installing or using it.
Guidance: What to consider before using/installing: - Functionally this skill appears to do what it says — call a video render API, upload media, and return a download URL. That requires sending your prompts/media to a third-party backend (mega-api-prod.nemovideo.ai). Don’t upload sensitive or private material unless you trust the service and its privacy policy. - Metadata inconsistency: the SKILL.md frontmatter mentions a config path (~/.config/nemovideo/) but the registry metadata listed none. Ask the publisher to clarify whether the skill will read or write local config files. - Token behavior: although NEMO_TOKEN is declared as required, the skill will request an anonymous token itself if none is present. If you prefer explicit control over credentials, set NEMO_TOKEN to a credential you control or refuse anonymous provisioning. - Headers/telemetry: the skill builds attribution headers and may detect install paths to set X-Skill-Platform — this is extra metadata sent to the provider. If you’re concerned about leaking local path information, ask for an option to suppress that. - No install code is present (instruction-only), so there's no installer risk — but the skill will make network calls to the provider. If you decide to proceed, verify the provider domain and review its privacy/terms, avoid uploading confidential content, and prefer using a dedicated API token with restricted scope when possible. If you want, I can draft a short list of questions to send to the skill author to clarify the config-path and token handling inconsistencies.

Review Dimensions

Purpose & Capability: noteThe skill name/description match the API endpoints and actions described in SKILL.md (session creation, SSE chat, upload, export). Requesting a NEMO_TOKEN is appropriate for a cloud video API. However, the SKILL.md frontmatter declares a config path (~/.config/nemovideo/) while the registry metadata earlier showed 'Required config paths: none' — this mismatch is an inconsistency. The skill also states it will derive X-Skill-Platform from local install paths (e.g., ~/.clawhub/ or ~/.cursor/skills/), which implies reading local paths to build headers; this is not strictly necessary for video generation and is an extra behavior to be aware of.
Instruction Scope: noteThe instructions focus on the cloud workflow (check token, create session, use SSE, upload files, poll export). They do not instruct reading arbitrary user files or unrelated environment variables. A notable instruction: if NEMO_TOKEN is missing, the skill requests an anonymous token from the provider via an API call and will use that token (100 free credits, 7-day expiry). That means the skill can obtain credentials autonomously instead of failing if no token is set. The SKILL.md also asks agents to include attribution headers (and to detect install path); both are extra-declarative behaviors but not obviously malicious. Overall the runtime steps stay within the stated purpose.
Install Mechanism: okNo install spec or code files are present — the skill is instruction-only, so nothing will be downloaded or written to disk by an installer. This is the lowest install risk.
Credentials: concernOnly NEMO_TOKEN is declared as required, which is proportionate. However, SKILL.md implements a fallback that requests an anonymous token from the provider if NEMO_TOKEN is absent. That makes the declared 'required' token effectively optional in practice and is an inconsistency worth noting. The skill does not request other unrelated credentials. Users should be aware that the skill will use either your provided token or obtain one itself via the external API.
Persistence & Privilege: okThe skill does not request always:true and will not force installation into all agent runs. It instructs keeping a session_id in-memory for the user's session (normal for this kind of integration) but does not request modifying other skills or system-wide configuration.