ClawHub

persona-simulator

v1.0.0

模拟自定义 AI 人格，使助手以特定性格、语言风格和行为特质与用户交互。当用户要求"扮演某种性格"、"以XX风格回答"、"模拟人格"、"角色扮演"时使用本 Skill。支持温和/直率/幽默/沉稳/俏皮/理性/感性等多种性格基调，以及自定义语言风格、行为特质和身份背景。

⭐ 0· 408·2 current·2 all-time
by@githubchenda
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
VirusTotalVirusTotal
Benign
View report →
OpenClawOpenClaw
Benign
high confidence
✓
Purpose & Capability
Name/description (persona simulation) match the content of SKILL.md and the included persona presets. The skill requires no binaries, no env vars, and has no install steps — all proportionate to an instruction-only behavior-modifier skill.
ℹ
Instruction Scope
SKILL.md constrains runtime behavior to follow a chosen persona, switch immediately on user request, avoid vulgar/violent content, and keep answers short. It also includes the rule '不解释设定、不自拆身份' (do not explain the persona setup / do not break character). This is coherent with roleplay goals, but the explicit directive to hide the persona/setup could enable non-transparent behavior if the agent uses the persona without an explicit user consent or when the user expects system transparency.
✓
Install Mechanism
No install spec and no code files beyond documentation — instruction-only. This is the lowest-risk model for disk/network footprint.
✓
Credentials
The skill requests no environment variables, credentials, or external config paths. All required data (persona descriptions) are shipped in the included reference file.
✓
Persistence & Privilege
always:false, user-invocable:true, and default autonomous invocation allowed — these are normal settings for a skill of this type. The skill does not request persistent installation or elevated privileges.
Assessment
This skill appears coherent and low-risk: it only provides persona instructions and includes no code, installs, or secrets. Before installing, consider whether you want an assistant that is explicitly instructed to 'not reveal the persona/setup' — that can reduce transparency and may surprise users who ask whether the assistant is roleplaying. If you require explicit transparency, update the SKILL.md to allow disclosure when asked. Also test the personas to ensure they respect safety boundaries in real prompts and do not escalate into disallowed content despite the stated bans.

Like a lobster shell, security has layers — review code before you run it.

latestvk97d4emsahhn25kap06gdj0jhs825ckt

License

MIT-0
Free to use, modify, and redistribute. No attribution required.
Termshttps://spdx.org/licenses/MIT-0.html

Comments