Skillv1.0.0

ClawScan security

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

Scanner verdict

BenignApr 29, 2026, 4:38 PM

Verdict: benign
Confidence: medium
Model: gpt-5-mini
Summary: The skill's requirements and runtime instructions are broadly consistent with a cloud video-from-audio service, with minor metadata and storage-related inconsistencies you should review before installing.
Guidance: This skill appears to do what it says: call the nemo video-rendering API to turn uploaded audio into synced videos. Before installing: 1) Verify the API host (mega-api-prod.nemovideo.ai) is a service you trust and read their privacy/retention policies — you will upload audio files there. 2) Decide whether you want the skill to auto-generate anonymous tokens (it will do so if NEMO_TOKEN is not provided); supplying your own token gives you more control. 3) Ask where the skill will persist the NEMO_TOKEN/session_id (disk vs memory) and ensure it won’t be printed or leaked. 4) Note the minor metadata inconsistencies (configPaths in SKILL.md but not in registry, and the declared required env vs auto-provisioning); these look like sloppy metadata rather than malicious behavior but you may want the publisher to clarify. If you plan to upload sensitive or copyrighted audio, confirm retention and access controls with the service first.

Review Dimensions

Purpose & Capability: okName/description match the instructions: the SKILL.md calls the nemo video rendering API to upload audio, create sessions, and render videos. Requesting a single API credential (NEMO_TOKEN) is appropriate. Minor inconsistency: the SKILL.md frontmatter lists a configPaths entry (~/.config/nemovideo/) that is not reflected in the registry summary; this is likely harmless but should be reconciled.
Instruction Scope: noteInstructions are focused on the stated task (auth, create session, upload file, SSE streaming, render/poll). They do instruct the agent to generate/store a client token/session_id and to detect an install path to set an attribution header; detecting install path implies reading filesystem locations (~/.clawhub/, ~/.cursor/skills/) which is outside pure API usage and should be limited to existence checks only. The skill explicitly tells the agent to avoid displaying raw tokens, which is good.
Install Mechanism: okNo install spec or code is provided (instruction-only), so nothing is written to disk by an installer. This is the lowest-risk install model.
Credentials: noteOnly one credential (NEMO_TOKEN) is declared, which fits a cloud API. However, SKILL.md will auto-request an anonymous token if NEMO_TOKEN is not present—this conflicts with the registry's 'required env var' assertion (declaring something as required but also auto-provisioning it is inconsistent). Consider whether you want the skill to auto-provision tokens on your behalf or prefer to supply your own token.
Persistence & Privilege: okSkill is not always-enabled and does not request elevated privileges. It will store a session_id / token for repeated use according to instructions; the storage location and lifetime are not specified, so confirm where and how session/token data are persisted (in-memory only vs. written to disk).