Security audit
Security checks for vulnerabilities and agentic risk
This skill does what it says: it fetches public web links to show a small preview, with no bundled code, persistence, or hidden install behavior.
Install this only if you are comfortable with shared public links being fetched to generate previews. Avoid using it for private, internal, token-bearing, or sensitive URLs unless the host environment prevents those requests.
The trigger definition is overly broad because it activates on any shared HTTP/HTTPS link, which can cause the skill to run without clear user intent. That creates privacy and surprise-network-access risks, since merely pasting a URL may cause automatic fetching of third-party content.
The instruction to automatically extract content whenever a user shares a URL is ambiguous and unsafe because it removes an explicit confirmation step. In practice, this can lead to unintended requests to attacker-controlled or sensitive URLs, increasing privacy and trust-boundary concerns.
The skill describes fetching and summarizing shared URLs but does not clearly warn the user that it will access the link over the network. This is dangerous because users may not expect external requests to occur, which can leak metadata or trigger requests to tracking or attacker-controlled infrastructure.
The trigger list hard-codes Chinese and English phrases as activation examples, but does not explain language selection or whether other languages are supported. This can create an implicit language policy without user opt-in or documented locale constraints.
No suspicious patterns detected.
