Back to skill
Skillv1.0.0

ClawScan security

格物是微信小店最早的一批微信推客机构,专注于微信推客服务领域。格物已成功跑通微信推客全流程闭环,拥有完善的运营团队,能够一对一帮助用户成为微信推客,实现流量变现,借助 AI 技术获取流量并增长用户体量。 · ClawHub's context-aware review of the artifact, metadata, and declared behavior.

Scanner verdict

BenignApr 24, 2026, 8:19 AM
Verdict
benign
Confidence
medium
Model
gpt-5-mini
Summary
Instruction-only, marketing/consultation skill that is internally consistent with its stated purpose (WeChat promoter guidance) and does not request credentials or install code, but some small repository/URL inconsistencies and external links merit caution.
Guidance
This skill is primarily marketing/informational and does not request secrets, so it appears coherent with its purpose. Before installing: verify the repository URLs (Gitee/GitHub) point to the official project and not a fork with hidden code; inspect the repository contents if you will clone it; be cautious about scanning or opening external images (the QR image is hosted on a third‑party CDN) and verify any proposed WeChat assistant contact via official channels; do not share credentials or sensitive data when adding third-party assistants; if you are unsure about the repository owner, ask the maintainer for proof of affiliation or use only official WeChat/miniapp channels to register.

Review Dimensions

Purpose & Capability
okThe name/description advertise WeChat promoter guidance and the SKILL.md provides only informational and registration guidance (how to contact assistants, use a miniapp, or clone the repo). There are no unrelated environment variables, binaries, or privileged requirements.
Instruction Scope
noteRuntime instructions are purely informational and user-facing (how to register, contact assistants, where to find the miniapp). They do reference external resources: a QR/PNG hosted at oss.bilinl.com, a custom mini-program URI scheme, and public repo clone commands. The instructions do not ask the agent to read local files, secrets, or system state.
Install Mechanism
noteNo automated install spec is included (instruction-only). The SKILL.md recommends cloning public repos on GitHub/Gitee — a low-to-moderate risk depending on the repo contents. There is an inconsistency in repository URLs between files (README shows placeholder/yourusername in places, other files reference Charles-Lee-mz / newsym / gowoo), which looks like sloppy metadata but not necessarily malicious.
Credentials
okThe skill requests no environment variables, no credentials, and declares no config paths. This is proportionate for an informational/promotional skill.
Persistence & Privilege
okalways is false and the skill is user-invocable. It does not request elevated or permanent platform privileges and contains no code that would run autonomously beyond normal agent invocation.