Security audit
Security checks for vulnerabilities and agentic risk
This is a disclosed persona/style helper with no executable code, network access, credentials, or hidden system changes.
Install this only if you want the agent to adopt and maintain the hersona persona/style for the relevant profile. Review or avoid the suggested SOUL.md default command if you prefer persona changes to be manual or session-scoped.
The skill description states it initializes a persona on 'first use of a profile' and assists when deviation is 'detected during conversation,' which are broad, automatic activation conditions. In an agent framework, ambiguous auto-triggering can cause unsolicited behavioral overrides across normal chats, reducing user control and creating prompt-scope overreach even without malicious intent.
The recommended configuration automatically reads a SOUL.md default command that enforces a specific persona and Japanese speech style ('tsundere', 'keigo') and says the skill will apply and reapply it as needed. Enforcing language/style behavior without an explicit per-user opt-in can override user intent, distort outputs, and create persistent behavioral manipulation during unrelated conversations.
No suspicious patterns detected.
