Skillv1.0.0

ClawScan security

Ai Video Pro Online · ClawHub's context-aware review of the artifact, metadata, and declared behavior.

Scanner verdict

BenignApr 19, 2026, 10:56 AM

Verdict: benign
Confidence: high
Model: gpt-5-mini
Summary: The skill's declared purpose (remote AI video editing) matches its runtime instructions and required credential (NEMO_TOKEN); there are a few minor metadata inconsistencies but nothing that indicates malicious intent.
Guidance: This skill appears to do what it says: it uploads your video to a third‑party cloud service (mega-api-prod.nemovideo.ai) for editing and returns a download URL. Before installing/using it: (1) confirm you are comfortable uploading your footage to that external domain and review any privacy/retention claims from the service; (2) be aware the skill uses a bearer token (NEMO_TOKEN) and will auto-request an anonymous token if none is present — anonymous tokens may have limited credits/expiry; (3) note the small metadata inconsistencies (declared configPaths vs registry metadata, and required env var vs runtime anonymous-token flow) — these are likely benign but worth knowing; (4) because this is an instruction-only skill that makes network requests, avoid putting long‑lived secret credentials into its environment unless you trust the remote service. If you want higher assurance, ask the publisher for a homepage or privacy policy and verify the domain and ownership before sending sensitive content.

Review Dimensions

Purpose & Capability: noteThe skill is an instruction-only connector to a cloud video-editing backend (mega-api-prod.nemovideo.ai) and only requests a single bearer token (NEMO_TOKEN) which is appropriate for that purpose. Minor inconsistency: the SKILL.md frontmatter lists a configPaths entry (~/.config/nemovideo/) while the registry metadata shows no required config paths; this is an implementation/metadata mismatch but does not change the core capability.
Instruction Scope: okSKILL.md instructions are focused on creating a session, uploading user-supplied media, sending edits via SSE, polling render status, and returning a download URL. It explicitly instructs not to leak tokens or raw API responses. It also instructs reading the skill's frontmatter and install path to populate attribution headers — that requires reading skill files/install path but is reasonable and scoped to attribution.
Install Mechanism: okThere is no install spec and no code files (instruction-only). That is the lowest-risk model; nothing is downloaded or written to disk by an install step.
Credentials: noteOnly a single credential (NEMO_TOKEN) is required, which fits a backend API. One minor contradiction: the manifest marks NEMO_TOKEN as required, yet runtime instructions include logic to obtain a short-lived anonymous token from /api/auth/anonymous-token if NEMO_TOKEN is absent. This is plausible design (allow anonymous usage) but is a behavioral mismatch between declared requirements and runtime behavior that users should be aware of.
Persistence & Privilege: okalways is false and the skill does not request persistent/privileged platform presence. It does instruct creating and using short-lived session tokens for render jobs; that is expected for a cloud rendering service.