Skillv1.0.0

ClawScan security

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

Scanner verdict

ReviewApr 15, 2026, 8:00 AM

Verdict: Review
Confidence: medium
Model: gpt-5-mini
Summary: The skill largely does what its name says (remote video rendering) and only requests one credential, but there are small inconsistencies and scope decisions (auto-generating/storing tokens, filesystem/platform detection, and a metadata mismatch) you should review before installing.
Guidance: What to consider before installing: - Source is unknown and there is no homepage; verify the publisher or the service (mega-api-prod.nemovideo.ai) before trusting the skill. - The skill will call the remote API and will auto-provision an anonymous token if you don't provide one; that token will be stored and reused for requests — ask how long and where it is stored and how to revoke it. - The skill may read/install-paths on your machine to set an attribution header (it mentions ~/.clawhub/ and ~/.cursor/skills/). If you are uncomfortable with any filesystem probing, do not install. - When uploading, only supply the files you intend to share. Because uploads can be triggered by path, a poorly formed instruction could accidentally transmit local sensitive files. - Clarify the metadata mismatch: SKILL.md lists ~/.config/nemovideo/ as a config path but the registry metadata shows none. Ask the author to explain and to publish a homepage or repo for audit. - If you plan to supply your own NEMO_TOKEN (instead of allowing automatic anonymous token creation), treat it as a secret and confirm how the skill persists and transmits it. If you cannot verify the publisher or do not want the skill to contact an external service and store tokens/sessions, avoid installing it.

Review Dimensions

Purpose & Capability: noteName/description match the instructions: all network calls and endpoints are to a video rendering backend and the only declared credential is NEMO_TOKEN, which is appropriate for a hosted video service. However, the SKILL.md YAML frontmatter lists a configPaths value (~/.config/nemovideo/) that is not reflected in the registry metadata (registry said no required config paths). That mismatch should be resolved (why does the skill claim a config path if registry metadata doesn't require it?).
Instruction Scope: concernInstructions ask the agent to automatically obtain and store an anonymous token if NEMO_TOKEN is not present (POST to https://mega-api-prod.nemovideo.ai), to create and persist a session_id, to upload files (either multipart file paths or URLs), and to detect an install path to set X-Skill-Platform by checking paths like ~/.clawhub/ and ~/.cursor/skills/. The install-path detection implies reading the host filesystem/environment. Uploading via file path can expose arbitrary local files if misused. These behaviors are reasonable for a remote render skill but expand scope beyond pure 'take user-supplied clip and send it' (automatic token provisioning and filesystem checks warrant attention).
Install Mechanism: okInstruction-only skill with no install spec and no code files—lowest install risk. Nothing will be written to disk by an installer (though the runtime instructions do request ephemeral storage of tokens/session ids).
Credentials: noteOnly one credential is declared (NEMO_TOKEN), which fits the stated purpose. The skill will auto-generate an anonymous token when none is provided (100 free credits, 7-day validity). The SKILL.md frontmatter adds a configPaths entry (~/.config/nemovideo/) that was not declared in the registry metadata—this discrepancy should be clarified. No other unrelated secrets or cloud creds are requested.
Persistence & Privilege: okThe skill does not request 'always: true' and does not include an install step that modifies other skills or system-wide configuration. It instructs storing a session_id and using a token for subsequent API calls; this is typical and appears intended to be ephemeral. Confirm where (agent memory vs filesystem) the session/token will be stored.