Security audit
Security checks for vulnerabilities and agentic risk
The skill is mostly coherent for fab equipment diagnostics, but it also exposes industrial data export and publish capabilities without clearly documented destination, approval, or scope controls.
Review this skill before production use. Install it only where the iaiops package is trusted, network access is segmented, allowed historian or UNS destinations are explicitly configured, and export or publish actions require the same approval and audit discipline as equipment writes.
The main instructional content begins in Chinese and continues throughout the skill description, which imposes a specific language/locale on users. The file does not indicate that Chinese is optional, nor does it offer an alternative language path or explicit user opt-in.
The skill repeatedly markets itself as read-first and MOC-gated for writes, but the documented toolset also exposes outbound data-moving functions such as historian_push, export_data, stream_publish, and uns_publish without equally explicit approval, destination, and scope controls in the manifest. In an industrial fab context, uncontrolled export or publish actions can leak production, recipe, alarm, or asset data to external systems and undermine operator trust in the documented safety model.
The manifest centers on SECS/GEM, OPC-UA, and cross-protocol operational analysis such as OEE, asset inventory, downtime RCA, and data quality. The additional capabilities compliance_mapping, compliance_frameworks, compliance_dengbao_levels, compliance_report, and compliance_evidence_bundle introduce a governance/compliance reporting function that is not clearly part of that stated mission. This appears to extend the skill into a separate assurance/reporting domain.
No suspicious patterns detected.
