social-persona-switcher

Guidance

This skill appears to do exactly what it says and asks for no extra permissions, so it is coherent from a security/privilege perspective. Before installing, consider the following: (1) Content safety — the persona styles include manipulative, sarcastic, or sexually/ gender‑laden language in examples (e.g., '女人', extreme flattery, passive‑aggressive sarcasm). That can be used to harass or socially manipulate others; decide whether you want the agent to generate such tones and consider adding guardrails or moderation (e.g., refuse targeted harassment, require consent). (2) Moderation behavior — the SKILL.md says to 'refuse overly negative input' but doesn't define criteria; if you need stricter safety, add explicit rules (block hate/abuse, PII, targeted harassment). (3) Monetization note — the doc mentions a paid “Pro” upgrade but provides no implementation; watch for later changes that add external endpoints or payment hooks. If you require stronger assurances (e.g., formal content filters, provenance logs, or limits on generating manipulative phrasing), request those additions from the publisher before enabling autonomous invocation.