Back to skill
Skillv1.0.0

ClawScan security

Ai Image To · ClawHub's context-aware review of the artifact, metadata, and declared behavior.

Scanner verdict

BenignApr 21, 2026, 3:46 AM
Verdict
benign
Confidence
medium
Model
gpt-5-mini
Summary
The skill's requirements and runtime instructions are coherent with an image→video cloud service: it needs a NEMO_TOKEN and talks to the nemovideo.ai API; nothing requested appears disproportionate, though there are a few small metadata/instruction inconsistencies you should be aware of.
Guidance
This skill appears to do what it says: it uploads your images and calls a third-party rendering API (mega-api-prod.nemovideo.ai) using a NEMO_TOKEN. Before installing, consider: (1) uploaded images and generated video will be sent to an external service — avoid sending sensitive images you wouldn't want stored/processed remotely; (2) the skill can create anonymous tokens itself if none are provided — verify you are comfortable with the service issuing ephemeral tokens and track how long they live; (3) confirm the domain and service legitimacy (nemovideo.ai) before granting or pasting any long-lived tokens; (4) the skill may probe local install paths to set an attribution header — if you object to any local filesystem probing, ask the author to remove that behavior or to explicitly declare the config path it will read; (5) consider using a throwaway/limited token rather than a full-account secret and revoke it when you stop using the skill.

Review Dimensions

Purpose & Capability
okThe skill describes a cloud video-rendering service and its SKILL.md only requires a single service token (NEMO_TOKEN) and makes API calls to a nemovideo.ai backend — this aligns with the declared purpose. Minor mismatch: registry metadata listed no required config paths, but the SKILL.md frontmatter includes a configPaths entry (~/.config/nemovideo/). This is likely benign but inconsistent.
Instruction Scope
noteInstructions are narrowly focused on session creation, file upload, SSE streaming, polling render status, and token renewal — all expected for a remote render service. The SKILL.md also instructs the agent to detect the install path to set an X-Skill-Platform header (checking ~/.clawhub/ and ~/.cursor/skills/). That implies reading local install paths (filesystem probing) which is outside the declared environment fields and should be noted.
Install Mechanism
okNo install spec or downloaded code — instruction-only skill. Lowest install risk (nothing is written to disk by an installer).
Credentials
noteOnly one credential is declared (NEMO_TOKEN), which is exactly what the service needs. The skill will auto-request an anonymous token via the public anonymous-token endpoint if NEMO_TOKEN is absent; that behavior is reasonable for a client but means the skill will make network calls to obtain tokens on first run. Also the SKILL.md references a config path (~/.config/nemovideo/) in its metadata even though the registry entry did not declare required config paths — a small inconsistency.
Persistence & Privilege
okThe skill is not marked always:true and does not request system-wide privileges. It needs to hold session_id and tokens for operations, which is normal for a session-based remote service; the SKILL.md does not instruct modifying other skills or system configs.