Skillv1.0.0

ClawScan security

Video Generation Like · ClawHub's context-aware review of the artifact, metadata, and declared behavior.

Scanner verdict

SuspiciousApr 29, 2026, 4:42 PM

Verdict: suspicious
Confidence: medium
Model: gpt-5-mini
Summary: The skill's instructions mostly match a remote video-generation service (it uploads files and uses a NEMO_TOKEN), but there are small inconsistencies and instructions that probe local config/install paths and persist tokens — actions that deserve scrutiny before you upload sensitive media or hand over credentials.
Guidance: Before installing or using this skill: (1) Understand it will upload the videos, audio, and images you provide to https://mega-api-prod.nemovideo.ai — do not send sensitive or private footage unless you trust that service and its privacy/retention policies. (2) The skill needs a NEMO_TOKEN; if you set one, make sure it's limited in scope and revokeable. If you don't provide a token, the skill will request an anonymous token from the API automatically (100 credits, 7-day expiry) — know that this still results in uploads to the remote service. (3) Ask the publisher to explain why the skill reads local install paths (~/.clawhub, ~/.cursor) and the skill frontmatter at runtime; those filesystem probes are not strictly necessary for generating videos and may reveal environment metadata. (4) Confirm the service's privacy/terms and where generated/ uploaded files are stored and how long they are retained. (5) If you need stronger assurance, request source/homepage or an auditable implementation (code) and a clear explanation of what is stored locally vs. remotely.

Review Dimensions

Purpose & Capability: noteName/description (generate videos matching a reference) aligns with the endpoints and workflow documented in SKILL.md. The required environment variable NEMO_TOKEN and the listed API endpoints fit the stated purpose. Minor inconsistency: the registry metadata supplied earlier said no required config paths, while the SKILL.md frontmatter declares a config path (~/.config/nemovideo/) — this mismatch should be clarified.
Instruction Scope: concernSKILL.md instructs the agent to call remote APIs, upload user-provided media, create and persist session tokens, and stream/poll job status — all expected for this feature. However, it also instructs the agent to read the skill's YAML frontmatter at runtime and detect the agent install path (e.g., check ~/.clawhub/ and ~/.cursor/) to populate X-Skill-Platform. Probing those filesystem locations and reading frontmatter is outside the core video-generation need and could reveal environment or install details; that needs justification or limits.
Install Mechanism: okNo install spec and no code files — instruction-only skill. No packages to download or archives to extract, which minimizes installation risk.
Credentials: noteThe skill requests a single credential (NEMO_TOKEN) which is appropriate for an API-backed video service. The SKILL.md also declares a config path (~/.config/nemovideo/) in frontmatter (useful for stored tokens) and asks the agent to generate an anonymous token if the env var is absent. These are plausible, but any token grants the ability to upload user files and start render jobs — confirm token scope/expiry and prefer revocable/limited tokens.
Persistence & Privilege: okalways:false (no forced global inclusion). The skill instructs saving session_id and tokens for session management, which is normal for a remote service. It does not request elevated system-wide privileges or modification of other skills' configs.