Security audit
Security checks for vulnerabilities and agentic risk
This skill is a disclosed pricing-data fetcher and calculator with limited, purpose-aligned network and local snapshot behavior.
Install only if you are comfortable with the skill contacting the configured pricing endpoints and updating its local latest_snapshot.json during full runs. Use --no-snapshot to avoid persistence, and do not treat stdout as one parseable JSON object without post-processing the labeled sections.
The skill claims to output standardized JSON, yet the documented behavior includes labeled prose sections, optional verification tables, and human-readable diff reports instead of one consistent JSON object. In automated pipelines, this mismatch can break downstream parsers, cause incorrect trust assumptions, and enable data handling errors when consumers treat mixed-format output as structured-only.
The skill claims to output standardized JSON, yet the documented behavior includes labeled prose sections, optional verification tables, and human-readable diff reports instead of one consistent JSON object. In automated pipelines, this mismatch can break downstream parsers, cause incorrect trust assumptions, and enable data handling errors when consumers treat mixed-format output as structured-only.
Without declared permissions the skill's intent is opaque and cannot be validated.
The skill says the output 'strictly follows this exact structure' and the examples use Chinese section labels such as 模型倍率, 模型补全倍率, 分组倍率, and 与上次数据对比. This appears to impose a specific language/locale on output without offering opt-in, alternatives, or justification for the locale requirement.
The skill specifies an exact output structure using Chinese section labels such as 模型倍率, and later shows a diff report in Chinese. This imposes a specific language on users without opt-in or documented regional justification, which is a natural-language policy concern.
The entire markdown file is written in Chinese and does not indicate that users may choose another language or that the content is intentionally limited to a Chinese-speaking or region-specific audience. Under the stated policy, forcing a specific language without user opt-in can constitute a natural-language policy violation.
The documentation says each run automatically overwrites data/latest_snapshot.json but does not prominently warn users that local files will be modified. This can lead to unintended data loss, corruption of previous baselines, or unsafe execution in environments where local state should remain immutable unless the user explicitly approves writes.
This markdown file documents that each run saves current data as data/latest_snapshot.json and overwrites the previous one. Because this affects local data persistence and replaces prior snapshot contents, the description should clearly warn users that historical snapshot data will be lost unless they preserve it elsewhere.
This markdown file contains user-facing instructions exclusively in Chinese, and there is no indication that users can opt into another language or that the skill is intentionally limited to a Chinese-speaking context. Under the language/locale policy rule, forcing a specific language without opt-in can be a natural-language policy violation.
SQP-3 applies to all file types and covers language or locale policy violations in natural-language content. The module docstring, usage text, CLI help strings, and status messages are presented exclusively in Chinese, with no opt-in language selection or indication that the skill is intentionally region-specific.
This code file contains natural-language text in the module docstring that is exclusively Chinese, including the usage instructions. Under the policy, forcing a specific language without user opt-in is a natural-language policy concern, and there is no indication that the tool is region-specific or that alternative locales are supported.
The argparse description and option help text are all presented in Chinese, so users are required to operate in that locale. The file does not offer a language selection mechanism or explain that the script is intentionally limited to a Chinese-speaking context.
No suspicious patterns detected.
