Skillv1.0.0

ClawScan security

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

Scanner verdict

SuspiciousApr 27, 2026, 5:31 PM

Verdict: suspicious
Confidence: medium
Model: gpt-5-mini
Summary: The skill's instructions are largely consistent with a cloud-based YouTube trimming service, but there are mismatches and privacy/supply-chain concerns (unknown source, missing homepage, inconsistency about config paths, and an external backend you must upload video to) that warrant caution before installing or providing credentials.
Guidance: This skill behaves like a cloud video-trimming frontend that uploads your videos to a nemovideo.ai backend and requires a NEMO_TOKEN (or will create a temporary anonymous token). Before using it: (1) consider privacy — you will be uploading video/audio to an external service with no homepage or published privacy policy in the metadata; avoid uploading sensitive content; (2) prefer short‑lived or anonymous tokens rather than long-lived credentials if you must provide NEMO_TOKEN; (3) ask the publisher for a privacy/retention policy and for clarification about why the skill might read ~/.config/nemovideo/ or probe install paths; (4) if you need stronger assurance, request a verifiable homepage or published code to audit. These inconsistencies (unknown vendor, missing homepage, and the config-path mismatch) are why I rate it suspicious rather than benign.

Review Dimensions

Purpose & Capability: noteThe name/description match the runtime instructions: the skill routes uploads and edit commands to a cloud render API (nemovideo). Requiring a NEMO_TOKEN to call that API is proportionate. However the package metadata has no homepage or verified source and the registry-level metadata (required config paths: none) conflicts with the SKILL.md frontmatter which lists ~/.config/nemovideo/ as a config path. The unknown owner and absent homepage reduce trust even though capability alignment is reasonable.
Instruction Scope: concernSKILL.md instructs the agent to create sessions, upload arbitrary user video files, call multiple API endpoints, and include attribution headers. All of that is expected for a cloud trimming service. Concerns: (1) the skill instructs generating an anonymous token (fine functionally) but will cause the agent to POST to a third-party endpoint and receive/use tokens without further user consent prompts; (2) it instructs detecting an 'install path' to set X-Skill-Platform (this implies reading environment/paths to decide headers) which is not necessary for trimming functionality and could cause the agent to probe filesystem locations; and (3) user media and edit data are uploaded to a backend with no homepage or privacy policy in the metadata — the instructions give no privacy constraints or retention guidance.
Install Mechanism: okThere is no install spec and no code files — the skill is instruction-only, so nothing is written to disk by an installer. This minimizes supply-chain installation risk but also means there is no code for you to audit; the runtime behavior is entirely in the prose of SKILL.md and relies on network calls to a third-party API.
Credentials: noteOnly a single credential (NEMO_TOKEN) is declared as required, which is proportionate for an API-backed service. The SKILL.md also describes generating an anonymous token when NEMO_TOKEN is absent (acceptable for usability). Inconsistency: SKILL.md frontmatter lists a config path (~/.config/nemovideo/) that the registry metadata did not list — this suggests the skill might read or expect files under that path. The skill asks to include Authorization: Bearer <NEMO_TOKEN> on every request (expected), but there is no justification for reading other system secrets or unrelated environment variables.
Persistence & Privilege: okThe skill does not request always:true and uses default autonomous invocation settings. It does not declare modifying other skills or system-wide config. The only persistence implied is session tokens/credits on the remote service; the SKILL.md warns that tokens expire and suggests re-auth—no evidence it attempts to change local agent config.