Security audit
Security checks for vulnerabilities and agentic risk
This is a disclosed Chinese K12 math practice skill that stores limited learning progress and can queue consented reminders, with no evidence of hidden code or data exfiltration.
This skill is reasonable to install for Chinese-language middle/high-school math practice. Before using it with minors or outside Mainland China, confirm local consent rules, emergency-contact localization, and whether profile sharing and reminders are enabled.
The manifest description at L003 states '本 SKILL 不自己提醒', implying the skill should not perform reminder behavior on its own. However, Module D explicitly says that upon student consent it will generate a reminder_enqueue request for weekly growth checks, which is operational reminder behavior initiated by this skill even if delivery is delegated to another skill.
The file title and all instructional content are written in Chinese, and there is no natural-language indication that users may choose another language or that the skill is intentionally limited to a Chinese-language or China-specific educational context. Under the stated policy, forcing a specific language without user opt-in can be a locale-policy violation.
Suspicious Unicode normalization or mixed-script content
The entire file is written in Chinese and presents the resource as the reference material for the skill, with no indication that users may choose another language or locale. Under the policy, forcing a specific language without user opt-in is a natural-language policy violation unless the locale constraint is explicitly documented and justified.
Suspicious Unicode normalization or mixed-script content
Suspicious Unicode normalization or mixed-script content
This file presents all operational safety instructions exclusively in Chinese, and it does not indicate that users or deployers can choose another language. Under the policy, forcing a specific language without user opt-in is a natural-language policy violation unless the locale constraint is clearly documented and justified.
The shared hint-ladder allows escalation to L5/L6, including a full walkthrough of the original problem, while this skill is explicitly a gradient trainer whose own shared table caps it at L4 and says to step back down when the layer is too hard. That mismatch can cause the agent to drift into prohibited in-the-moment problem solving, weakening intended guardrails around practice-only behavior and potentially violating the skill’s routing boundary to the dedicated problem-solving coach.
The file hard-codes the skill library for Mainland China Chinese K12 use and says deployments to other regions should not be used until localization work is done. In an education product used by minors, this can cause unsafe or inappropriate crisis contacts, legal-consent assumptions, and curriculum guidance if the user is outside the intended region, especially when no user choice or runtime locale confirmation is required.
This markdown file is written entirely in Chinese and establishes repository-wide terminology and normative wording, but it does not indicate that language choice is optional or limited to a Chinese-only deployment context. Because the policy category applies to natural-language content in any file type, this is a locale/language constraint that appears to be imposed without user opt-in.
No suspicious patterns detected.
