Security audit
Security checks for vulnerabilities and agentic risk
The skill is an HTML checklist, but it asks for command/write capability and triggers on messaging/email tasks that do not match that purpose.
Review before installing. The HTML advice itself is ordinary, but only install it if you are comfortable giving this skill write and command-execution authority and ignoring its mismatched messaging/email trigger language.
Hidden instructions were detected in comments or invisible text. These could contain malicious directives. Manual review is recommended.
<!-- 外链必加 noopener noreferrer -->
<a href="https://example.com" target="_blank" rel="noopener noreferrer">外链</a>
<!-- 用户生成内容加 nofollow ugc -->
<a href="https://user-site.com" rel="nofollow ugc">用户链接</a>
Hidden instructions were detected in comments or invisible text. These could contain malicious directives. Manual review is recommended.
<!-- 模态用 dialog：内置焦点陷阱与 ESC -->
<dialog open>
<p>内容</p>
<form method="dialog"><button>关闭</button></form>
Hidden instructions were detected in comments or invisible text. These could contain malicious directives. Manual review is recommended.
<!-- 模态用 dialog:内置焦点陷阱与 ESC -->
<dialog open>
<p>内容</p>
<form method="dialog"><button>关闭</button></form>
The trigger condition says to use the skill for message sending, notification push, email/SMS, and communication integration, which is unrelated to the skill's stated HTML accessibility/SEO purpose. Overbroad and mismatched triggers can cause the agent to invoke this skill in inappropriate contexts, potentially exposing exec/write capabilities during workflows that should never need them.
The skill advertises create/modify/delete/save/export style operations via input and output parameters but does not clearly warn that files or artifacts may be changed. In an agent setting, undocumented write semantics are risky because users may believe they are requesting analysis only while the skill is authorized to alter content.
The skill presents itself as a pure Markdown-based HTML self-check tool with no extra API requirements, but later introduces MD+EXEC behavior and network-diagnostic command usage such as ping. This creates a capability mismatch that can mislead users and agents into permitting command execution in contexts where only passive linting was expected, increasing the chance of unnecessary system or network actions.
The skill includes command-execution and network-diagnostic behavior that is not justified by its stated purpose of HTML quality checking. Unnecessary execution primitives expand the attack surface and could be abused by downstream prompts or users to run local commands or probe connectivity under the guise of a benign HTML tool.
No suspicious patterns detected.