Back to skill
Skillv1.0.0
ClawScan security
create-virtual-girlfriend(虚拟女友) · ClawHub's context-aware review of the artifact, metadata, and declared behavior.
Scanner verdict
BenignApr 18, 2026, 3:42 PM
- Verdict
- Benign
- Confidence
- medium
- Model
- gpt-5-mini
- Summary
- The skill's files and runtime instructions are consistent with a virtual companion chatbot: no installs, no credentials, and no unrelated capabilities — but the memory/persistence behavior is unspecified, so confirm privacy details before use.
- Guidance
- This skill appears coherent and matches its stated purpose, but before installing or using it consider: (1) Where are 'memories' stored (agent memory, local files, remote DB)? (2) Who can access those saved memories and for how long? Is there a way to view/export/delete them? (3) Confirm that the agent's Read/Write/Edit tool usage is sandboxed and doesn't expose other files or secrets. (4) Do not share passwords, credit cards, health details, or other sensitive info in chats. If possible, ask the skill publisher for a privacy/retention statement and a clear way to delete stored data.
Review Dimensions
- Purpose & Capability
- okThe name/description (virtual girlfriend / emotional companion) align with the provided content: personalities, conversation examples, and a memory guide. The skill does not request unrelated binaries, credentials, or config paths.
- Instruction Scope
- noteSKILL.md contains only conversational guidance and a memory guide (what the assistant should remember). It does not instruct the agent to access external endpoints, credentials, or arbitrary system files, but it does assume a memory/persistence mechanism without specifying where or how memories are stored.
- Install Mechanism
- okNo install spec and no code files — instruction-only skill. Nothing is downloaded or written to disk as part of an installer.
- Credentials
- okThe skill requests no environment variables, credentials, or config paths. There are no obvious requests for unrelated secrets or services.
- Persistence & Privilege
- notealways:false and autonomous invocation is allowed (platform default). The memory-guide implies persistent storage of user preferences and conversation history; the SKILL.md does not declare where memory is kept, retention policy, or how to delete data. Allowed-tools include Read/Write/Edit which could be used to read/write agent-local memory or files — this is reasonable for personalization but should be clarified.