Security audit
Security checks for vulnerabilities and agentic risk
This HR skill is mostly coherent, but it combines broad activation triggers with write, network, session, and subagent authority for high-impact agent lifecycle actions.
Review this skill before installing in any environment where agents can modify records, sync private knowledge bases, send session messages, spawn subagents, or delete model/data assets. Use it only with explicit approval gates for retirement, deletion, permission assignment, and external API actions, and consider narrowing its activation triggers.
The trigger list is broad, multilingual, and includes generic terms like "HR," "ethics," and lifecycle-related phrases, which can cause the skill to activate in contexts where the user did not intend to invoke it. Because the skill has write, network, and subagent/session permissions, unintended activation can lead to downstream sensitive actions or data handling beyond user expectations.
The skill advertises HR lifecycle actions including onboarding, assessment, and retirement, while the manifest grants file write and network/API permissions, yet the description does not clearly warn users about these capabilities or the possibility of destructive retirement-related operations. In an HR/governance context, hidden state-changing behavior is especially risky because it may modify records, propagate decisions externally, or initiate deletion/retirement workflows without informed consent.
The skill content and interface descriptions are presented in a mixed Chinese-English format, and trigger phrases are defined in specific languages without stating whether users may interact in other languages or choose a preferred locale. This can amount to an implicit language policy constraint that is not documented as optional or justified.
No suspicious patterns detected.
