Security audit
Security checks for vulnerabilities and agentic risk
This skill is a disclosed UIBC education-domain judging aid with a local read-only checker and no evidence of credential access, network use, persistence, or destructive behavior.
Install only if you want UIBC/TH-EDU-001 education-domain judging support. Be aware that broad trigger words may activate it in ordinary education or accountability discussions, so confirm the UIBC EDU context before relying on its output.
The trigger conditions in metadata are broad enough that ordinary mentions of 'education', 'UIBC', 'evidence package', or 'responsibility attribution' could invoke the skill outside its intended niche. Over-broad activation can cause unintended routing, making the agent apply specialized judging logic in irrelevant contexts and potentially produce misleading authoritative outputs.
The 'when to use' section repeats ambiguous activation criteria like generic references to education, migration judgment, evidence packages, and accountability, without strong boundaries. This increases the chance of accidental invocation during normal discussion, which can lead to context confusion, incorrect tool selection, and overreach by a domain-specific skill presenting itself as applicable authority.
The trigger list includes several broad terms such as "教育", "证据包", and "责任归属" that are not unique to this specific skill and could match many unrelated governance, education, or policy contexts. That raises the chance of unintended invocation, which can cause the wrong skill to activate, inject irrelevant instructions into another workflow, or interfere with higher-priority safety-relevant routing.
