Security audit
Security checks for vulnerabilities and agentic risk
This skill is mainly a Chinese social-platform hot-topic ranking tool, with some broad auto-run behavior and extra content-creation templates users should understand before installing.
Install this if you want Chinese social-platform trend monitoring and ranking. Be aware it is designed to run automatically for broad hot-topic prompts and may use network access when available, so use explicit prompts if you want tighter scope. Treat bundled content-creation templates as optional reference material and verify sources, copyright, advertising, AIGC, and platform-compliance claims before publishing anything based on its output.
The skill description advertises default automatic execution for broad requests like checking what's hot, which overlaps with many normal user queries. This can cause the agent to invoke a networked or data-processing skill without clear user intent, leading to over-collection, unexpected external access, or surprising behavior.
The instruction to 'directly start scanning' without asking clarifying questions for vague inputs removes an important safety checkpoint. In practice, this makes accidental invocation more likely and can trigger resource use, data access, or network retrieval in contexts where the user did not actually request such an operation.
The user input mapping treats generic everyday phrases such as '找话题' and '抓热点' as direct execution triggers. Because these phrases are common and context-dependent, the skill may activate unintentionally during ordinary conversation, causing unauthorized or unexpected scanning behavior.
The file’s content is materially unrelated to the declared hotspot-tracking purpose and instead provides a full short-video voiceover production template. This kind of scope drift is dangerous because it can cause the agent to generate persuasive distribution content rather than objective monitoring output, increasing the risk of misuse for influence operations, spammy amplification, or deceptive repurposing of scraped trends.
The skill appears capable of reading local reference files and potentially writing data, but it declares no explicit tool scope or permission boundaries. In an agent environment, missing scope declarations can let the skill access filesystem capabilities more broadly than users or platform policy expect, increasing the chance of unintended data exposure or side effects.
The description contains very broad activation phrases such as '最近什么最热' and '帮我抓全网热点', plus '默认自动执行，不先反问用户', which can cause the skill to trigger on common, ambiguous requests without clear user consent. In an agent setting, this increases the chance of unintended invocation, unnecessary external data access, and content-routing mistakes, especially because the skill is designed to scan multiple social platforms automatically.
The description is written to enforce Chinese-language behavior and Chinese-platform assumptions without stating that this is optional or user-selectable. This can cause incorrect routing, user confusion, and reduced transparency when users communicate in other languages or expect locale-neutral behavior, though it is less severe than a direct security flaw.
The template explicitly instructs automatic execution and says not to ask follow-up questions, which can cause the skill to trigger on ambiguous requests and act without confirming user intent. In a cross-platform hot-topic scraping skill, this increases the chance of unintended invocation, over-collection, and outputs based on inferred rather than user-provided constraints.
This markdown template is entirely specified in Chinese and presents the reporting format as the default output format, with no indication that users may choose another language. That creates a natural-language locale constraint without opt-in, which matches the policy-violation category for language or locale restrictions.
The guidance goes beyond analytics and teaches influencer-style scripting techniques such as hooks, pain points, engagement prompts, and conversion-oriented calls to action. In the context of a hotspot-monitoring skill, this creates a dangerous bridge from detecting trending topics to mass-producing persuasive content about them, which can facilitate manipulation, astroturfing, or low-transparency promotion.
This markdown file contains all user-facing instructions in Chinese and does not indicate that users can select another language or that the skill is intentionally limited to a Chinese-speaking audience. Under the policy rule, forcing a specific language without opt-in is a natural-language policy violation.
This markdown file presents all guidance, headings, and examples exclusively in Chinese, which effectively forces a specific language on users. Under the policy, language constraints should either offer user choice or be clearly documented as justified for a region-specific tool.
The entire skill document is written in Chinese and provides no indication that other languages are supported, no opt-in language selection, and no justification that the skill is region-specific. Under the policy, forcing a specific language or locale without user opt-in is a natural-language policy violation.
The template title and instructions are written to produce output in Chinese by default, with no indication that users may choose another language or locale. This is a natural-language policy issue because it imposes a specific language without opt-in or justification.
The entire document is written in Chinese and does not indicate any user-selectable language or locale option. Under the policy, forcing a specific language without opt-in can be a natural-language policy violation unless the locale restriction is explicitly documented and justified.
The section introduces explicit trend-jacking and content-production guidance ('紧急蹭热度', output within fixed windows), which expands the skill from objective hotspot monitoring into persuasive operational advice. In this skill context, that creates scope drift: an auto-executing monitoring skill could steer users toward amplification tactics rather than neutral analysis, increasing misuse risk and undermining the stated objective of objective ranking.
The lifecycle section maps topic phases to recommended engagement windows and prioritizes when to capitalize on attention, which goes beyond monitoring into tactical exploitation of public attention. Because the skill is described as automatically executing without first asking the user, this context makes the expansion more dangerous by normalizing unsolicited strategic guidance rather than objective reporting.
The application guidance explicitly recommends content creation actions, production timing, and creator response strategies, which materially exceeds the stated purpose of objective cross-platform hotspot scanning and denoised ranking. This is dangerous because it can transform a discovery tool into an influence-optimization assistant, enabling coordinated amplification, sensationalism, or manipulation around trending topics.
This code embeds user-facing platform labels in Chinese and later emits multiple Chinese-language report fields and messages. Because the skill provides no language selection or documented locale constraint, it effectively forces a specific language on users, which is a natural-language policy violation under the locale-choice rule.
The entire template is written as a Chinese-only output format and provides no indication that users may choose another language or locale. For skills intended for general use, forcing a specific language without user opt-in can violate language/locale policy expectations.
This markdown file consistently instructs users in Chinese and does not indicate that the language is optional or selectable. Under the policy for natural-language violations, forcing a specific language without user opt-in can be a locale-policy issue.
No suspicious patterns detected.
