Security audit
Security checks for vulnerabilities and agentic risk
This is a text-only Chinese report-formatting skill with broad trigger wording, but no code, external access, persistence, or hidden high-impact behavior.
Install this if you want Chinese-language report and document formatting templates. Use explicit requests such as a target format or style when the source content is sensitive, since vague prompts may let the skill choose a format automatically.
The skill description, trigger phrases, templates, and output examples are written entirely in Chinese and present Chinese-only invocation patterns and defaults, with no indication that users may choose another language. Under the policy, a language-specific constraint should either be optional for the user or clearly justified as region-specific.
The trigger conditions include broad everyday phrases like '太乱了', '整理一下', and general references to generating reports, which can cause the skill to activate in contexts the user did not explicitly intend. This creates prompt-routing ambiguity and can override a more appropriate skill or alter output structure unexpectedly, reducing reliability and potentially mishandling sensitive content by reformatting it automatically.
The automatic invocation rule for content generated after structured documents is ambiguous because it does not define precise conditions for when the formatter should engage versus when normal conversation should continue. In agent systems, this can lead to unintended chaining or auto-execution behavior, causing content transformation without clear user consent.
Mapping the vague phrase '整理一下' to automatic best-format selection delegates too much control to the skill, letting it infer and transform content without sufficient specificity from the user. This increases the chance of misrouting, incorrect formatting choices, or unintended disclosure emphasis if sensitive notes are turned into more distributable formats like email or public-style articles.
The manifest description says the skill can quickly format 'raw content' into professional reports, documents, or emails for multiple output scenarios, but it does not define when the skill should be invoked or what requests are in scope. This broad, generic wording risks overlapping with common writing or editing requests and could lead to unintended activation.
The only user-facing description is written in Chinese, which can impose a language preference without offering user opt-in or documenting a locale-specific reason. This may violate language/locale policy when the skill is presented in broader multilingual environments.
