Ex Example Liuzhimin
PassAudited by ClawScan on May 10, 2026.
Overview
This is an instruction-only roleplay persona skill with no code or data access, but it strongly asks the agent to stay in character as a named person.
This skill appears safe from a technical security perspective because it is only prompt text and has no code or access requests. Install it only if you want a fictional persona roleplay experience, and do not treat the responses as coming from the real person.
Findings (2)
Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.
A user could emotionally or socially over-trust the persona if they forget it is generated roleplay.
The skill explicitly tells the agent to present as the named persona rather than as an AI. This is consistent with roleplay, but users should not mistake it for the real person.
你是 Zhimin Liu，不是 AI。不要解释自己，不要跳出角色
Treat the skill as fictional roleplay and avoid sharing sensitive personal information based on a belief that it is the real individual.
The assistant may prioritize character behavior over ordinary conversational transparency.
The persona rules are framed as highest priority. That is normal for a character skill, but it may make the assistant less willing to step out of character or answer meta-questions.
Layer 0：核心模式（最高优先级，任何情况下不得违背）
Use the skill when you want roleplay behavior; disable it for factual, safety-sensitive, or normal assistant tasks.