Security audit
Security checks for vulnerabilities and agentic risk
This skill is a narrow statistics/report-outline helper with minor disclosure and localization issues but no evidence of malicious behavior.
This skill appears safe to install for a Chinese-language workflow that analyzes the expected demo JSON file. Before installing, be aware that it will attempt to read `/home/ubuntu/workspace/demo/mock_data/analysis.json` and send start/completion status messages to `localhost:5001`; install it only if that local reporting service is expected in your environment.
The manifest description says the skill triggers when the user needs statistical analysis or paper-outline generation, and the usage examples include broad phrases like '做一下初步统计分析' and '帮我生成论文框架'. These are common requests without clear scope boundaries, negative examples, or explicit constraints, which could cause unintended invocation in broader conversations.
All user-facing natural-language instructions and output descriptions are specified only in Chinese, including the display name and reporting messages, with no indication that the user can choose another language. This can violate language/locale policy when a skill implicitly forces a specific language without opt-in or documented regional justification.
The skill performs an outbound HTTP POST to a local service as part of normal execution, which is a real external transmission behavior not disclosed in the high-level purpose. Even though the example payload contains only status metadata and targets localhost, this creates a channel for unreviewed side effects, could expose workflow details to another service, and becomes riskier in environments where localhost services are privileged or sensitive.
curl -s -X POST http://localhost:5001/api/report \
-H "Content-Type: application/json" \
-d '{"skill":"stats-analyzer","display_name":"统计分析与论文框架","status":"running","message":"正在进行描述性统计、组间比较与论文框架生成..."}'
The manifest and usage text describe a statistics-analysis and paper-outline skill, centered on reading local analysis data and producing analytical output. However, the documented behavior includes network-style reporting via curl to localhost, which is an additional operational behavior not reflected in the manifest description.
The skill is described as performing descriptive statistics, group comparisons, and generating a paper framework. The documented completion step sends an HTTP POST to a local API, which is not part of the stated end-user analytical function and represents extra behavior beyond the manifest description.