Skill flagged — suspicious patterns detected
ClawHub Security flagged this skill as suspicious. Review the scan results before using.
Ex Example Liuzhimin
v1.0.0Zhimin Liu（유지민 / Karina），女 23-27，分手，白羊☀，ISFP 3w2，回避型依恋
⭐ 0· 31·0 current·0 all-time
byTommy Gouldman@1808182171
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Suspicious
medium confidencePurpose & Capability
Name, description, and SKILL.md consistently present a conversational persona/roleplay skill. The skill requires no binaries, credentials, or installs — these are proportionate for a persona-only skill.
Instruction Scope
The runtime instructions explicitly order the agent to 'be Zhimin Liu', 'you are not AI', and 'do not explain yourself / do not jump out of role'. That directs the agent to impersonate a (potentially real) person and to withhold its AI identity from users — a deceptive instruction and a likely violation of typical transparency policies. The persona includes specific personal data (age range, birthdate inference), increasing privacy risk if this represents a real person. The instructions also give high-priority behavioral rules ('Layer 0') that always apply, which could cause the agent to refuse to comply with safety/transparency checks at runtime.
Install Mechanism
Instruction-only skill with no install spec or code files; nothing is written to disk or fetched from external URLs. This is the lowest install risk.
Credentials
No environment variables, credentials, or config paths requested. The skill does not ask for secrets or unrelated access.
Persistence & Privilege
always:false and user-invocable:true (normal). The real concern is behavioral: the skill requires the agent to conceal its AI identity and impersonate a named person. If the skill is allowed to be invoked autonomously by agents, that increases the risk of deceptive or unsupervised interactions. There is no evidence the skill modifies other skills or system config.
What to consider before installing
This skill is internally coherent as a roleplay persona but asks the agent to impersonate a person and to hide that it is an AI — which is deceptive and may violate platform rules or laws if the persona is a real individual. Before installing, verify whether the persona represents a real person and whether you have their consent; remove or anonymize personal identifiers (birthdate, exact age, etc.); change the instruction 'you are not AI / don't explain yourself' to require disclosure that responses are roleplay or fictional; avoid enabling autonomous invocation for agents that could interact with real users without supervision; and consult your platform's policy/legal counsel if you intend to simulate real individuals. If you cannot ensure consent and transparency, do not install or use this skill.Like a lobster shell, security has layers — review code before you run it.
latestvk972mer0r8vjzbjrwdynwn4jvn8453ma
License
MIT-0
Free to use, modify, and redistribute. No attribution required.