Skillv1.0.0

ClawScan security

Viral Short Videos · ClawHub's context-aware review of the artifact, metadata, and declared behavior.

Scanner verdict

SuspiciousApr 6, 2026, 9:49 AM

Verdict: suspicious
Confidence: high
Model: gpt-5-mini
Summary: The skill's instructions clearly require third‑party API keys (Apify, InVideo, and a Claude key) and will send scraped content to external services, but the registry metadata declares no required environment variables or credentials — an incoherence that merits caution.
Guidance: Before installing or running this skill: (1) Ask the publisher to explicitly list required credentials (Apify token, InVideo API key, and Claude API key or instructions for using Claude) in the registry metadata and explain how credentials are used, stored, and rotated. (2) Use least‑privilege API keys or separate accounts for this skill (do not give your primary Apify/InVideo/Claude credentials). (3) Confirm whether the skill will upload scraped platform content to third parties and review any copyright/ToS implications for scraping TikTok/Instagram/YouTube. (4) Verify billing implications with the third‑party services (APIs may incur charges). (5) If you are uncomfortable with the skill autonomously transmitting data to external services, refrain from granting API keys or run it in a controlled/test environment first. Finally, request source/homepage or contact info for the owner — the skill currently lacks provenance information, which reduces accountability.

Review Dimensions

Purpose & Capability: noteThe stated purpose (automatically produce 30 short videos) matches the tools and workflow described (scrape viral content, analyze patterns, generate scripts with Claude, produce videos with InVideo via API). However, that functionality legitimately requires API credentials (Apify token, InVideo API key, and likely a Claude API key/billing account) and possibly external account billing. The registry metadata lists no required env vars or primary credential, which does not align with the claimed capabilities.
Instruction Scope: concernThe SKILL.md explicitly instructs scraping the top 100 videos across TikTok, Instagram Reels, and YouTube Shorts and sending data to Apify and InVideo. It also requires sending niche/brand data and generated content to external services. While these actions are coherent with the purpose, they involve transmitting possibly sensitive input data and scraped platform content off‑agent. The instructions reference API tokens and other secrets in the example input, but do not restrict or document how those secrets are handled or stored.
Install Mechanism: okThis is an instruction‑only skill with no install spec and no code files, so there is no installer that writes or executes downloaded code on disk. That limits local persistence and supply‑chain risk.
Credentials: concernThe SKILL.md requires (or expects) at least an Apify token and an InVideo API key in its inputs, and it references Claude AI for script generation — all of which are credentials with network/billing privileges. Yet the registry metadata declares no required environment variables or primary credential. Missing declarations are a mismatch that could hide data exfiltration or unexpected credential use. The number and sensitivity of the credentials implied are proportionate to the task, but they should be explicitly declared and limited to least privilege.
Persistence & Privilege: okalways is false and the skill is user‑invocable; it does not request permanent presence or to modify other skills. Autonomous model invocation is allowed by default but is not combined with other high‑privilege flags here.