Security audit
Security checks for vulnerabilities and agentic risk
This skill is a Chinese social-media content drafting tool with expected local output files and limited, disclosed-enough topic search behavior, with no evidence of hidden persistence, credential use, destructive actions, or exfiltration.
Install this as a drafting aid, not as an automated social-media operator. Review generated claims before publishing, and treat the trending-search feature as potentially sending your search terms to external Chinese web platforms if used.
The skill claims real-time trending-topic search and appears to rely on external network requests, but this behavior is not clearly declared and may not match actual implementation. Undeclared outbound requests can leak user prompts or metadata to third parties, and misleading capability claims reduce user ability to assess privacy and trust boundaries.
The skill advertises capabilities that imply file writing and network access but does not declare any explicit tool scope or permissions. This creates an authorization and transparency gap: a host may expose more capability than intended, or users may be unaware that external requests and file exports are possible.
The skill is presented primarily in Chinese and targets Chinese platforms, but it does not state that language behavior is optional or user-selectable. This can create a locale/language policy issue if the broader environment expects language choice rather than a forced default.
The trigger phrases are broad enough to activate during ordinary conversation about writing or social media, increasing the chance of unintended invocation. Over-broad activation can cause the skill to take over unrelated tasks, potentially invoking file or network-backed behavior without clear user intent.
The description is entirely in Chinese and presents the skill as a Chinese-language content creation assistant, with no indication that users can choose another language or locale. This can violate language/locale policy when a skill implicitly constrains interaction language without explicit opt-in or justification.
The platform configuration and generated content are built around Chinese platforms and Chinese-language day labels such as '周二' and '周四'. Across the file, output templates are written only in Chinese, but there is no opt-in, language selection, or documentation that this skill is intentionally limited to a Chinese-language audience.
The manifest describes real-time hot topic search as a core capability, which implies retrieving and presenting actual trending topics. This function merely issues a GET to a search page and, on success, returns metadata with a note telling the user to visit the URL manually, so the implemented behavior falls short of the claimed functionality.
This JSON template is entirely structured around Chinese-language field values and output strings, including section names, requirements, caption rules, and the script template. Because the file does not indicate that Chinese is optional or that the skill is explicitly limited to a Chinese-language/regional use case, it appears to impose a specific language by default.
The template uses Chinese-only section names, examples, and content guidance throughout the file, effectively enforcing a specific language/locale. The policy allows locale constraints only when they are clearly documented and justified or when the user is given a choice, neither of which appears here.
This JSON template uses Chinese section names and writing rules throughout, effectively constraining the skill output to Chinese. Under the policy, forcing a specific language without user opt-in or a clearly documented region-specific justification is a natural-language policy violation.
The skill says it activates when the user says "热点选题" or "给我一个选题", but "给我一个选题" is a very general request that could arise in many contexts unrelated to this skill. There is no explicit limitation tying this trigger to social-media content creation scenarios only.
The description presents automatic scheduling as part of the one-stop solution, suggesting posts may be scheduled operationally. In code, the function only creates calendar entries with suggested dates, captions, and prompts; there is no integration with platform APIs or any mechanism to submit scheduled posts.
No suspicious patterns detected.
