Security audit
Security checks for vulnerabilities and agentic risk
This skill is a narrow Gmail job-reply helper that accesses private email for its stated purpose, with no hidden execution, persistence, or exfiltration behavior found.
Install only if you are comfortable letting the agent use your authenticated gog Gmail account to search recent inbox and sent messages for job-related emails. Confirm the Gmail account and search scope before use, especially on shared or work machines.
The top-level description contains broad phrases like reading recruiter emails and checking inbox replies, which can trigger the skill on loosely related requests. Because the skill reads private Gmail content, unintended invocation can expose sensitive email data or cause the agent to access more mailbox content than the user explicitly meant to share.
The skill is designed to read and summarize Gmail inbox contents, including recruiter correspondence, but it does not prominently warn that this involves accessing private email data. Without an explicit privacy notice and consent checkpoint, users may not realize the extent of mailbox inspection, increasing the risk of unintended disclosure of sensitive personal or professional information.
The activation examples use vague natural-language triggers like 'check my job application replies' and 'read recruiter emails' without requiring explicit scope, account selection, or confirmation. In a mailbox-reading skill, ambiguous trigger phrases increase the chance of accidental activation and privacy-invasive retrieval of sensitive correspondence.
The activation examples include both Chinese and English phrases, but the document does not explain whether the skill intentionally supports both languages, how language is chosen, or whether outputs will follow the user's preferred locale. This can create an implicit language/locale behavior without explicit user choice.
No suspicious patterns detected.
