Security audit
Security checks for vulnerabilities and agentic risk
This is a small Markdown-only learning skill with broad Chinese triggers, but no code execution, credential access, persistence, or hidden high-impact behavior.
Install only if you want a Chinese-language learning note about collaborative agents and related videos. Be aware that its broad triggers may activate on general collaborative-agent help requests.
The skill metadata, description, and trigger conditions are written in Chinese and appear to steer activation and interaction in that language without explicit user opt-in. This can create consent and usability issues, cause mismatches between user language and skill behavior, and reduce transparency about what the skill does when auto-triggered.
The trigger phrase at this location is broad enough to activate on generic requests for related help rather than a clearly scoped invocation. That can cause unintended skill activation, overriding user intent or injecting irrelevant instructions into normal conversations; while not directly code-execution, it is a prompt-routing and safety-boundary weakness.
The merged secondary section introduces another vague trigger phrase, increasing the chance this skill fires when a user mentions a broad topic rather than intending to invoke the skill. Because the file auto-merges content from another source, ambiguous triggers also compound routing confusion and make accidental activation more likely.
