Security audit
Security checks for vulnerabilities and agentic risk
This email-writing skill is purpose-aligned and shows no evidence of exfiltration, hidden command execution, or unsafe system changes, though its local email-record code and broad triggers deserve user awareness.
Install only if you are comfortable with a Chinese-first email helper that may use broad email prompts and can read/write local JSON files. Avoid placing confidential email bodies into any saved-record workflow unless you have reviewed where sent_emails.json is stored and how to delete it.
The README presents the skill name, description, and usage examples entirely in Chinese, with no indication that other languages are supported or that the user may choose their preferred language. This can violate language/locale policy when a skill implicitly constrains interaction language without explicit opt-in or justification.
The example trigger phrases are very broad, natural-language requests that overlap with ordinary user intents such as writing leave, business, or thank-you emails. In agent ecosystems that auto-route skills based on prompt similarity, this can cause the skill to activate unexpectedly and capture requests that were not explicitly intended for it, increasing the risk of misrouting or unwanted prompt handling.
The trigger phrases are very broad and match common, everyday language such as writing or replying to emails. This can cause the skill to activate in contexts the user did not explicitly intend, increasing the chance of inappropriate handling of sensitive email content or unintended access to local template/storage features.
The skill requests file read/write access for template storage but does not clearly disclose to users that email templates or potentially sensitive drafted content may be stored locally. In an email-assistance context, drafts can contain personal, workplace, or confidential information, so undisclosed persistence creates a meaningful privacy and data-retention risk.
The module docstring explicitly states support for Chinese, English, and Japanese. In code, language handling in generate_email and main only supports zh/en templates and detection; there are no Japanese templates or logic, so the documentation contradicts actual behavior.
The skill's stated purpose is email drafting/refinement, but it also includes functionality to persist sent-email data locally and expose aggregate usage statistics. Even though the provided main flow does not clearly invoke the save path, retaining email records can capture sensitive business or personal content and metadata without clear user consent or disclosure, creating a privacy and data-minimization issue.
Writing email content to a local JSON file without an explicit user-facing disclosure at the write point is a legitimate privacy/security concern. Emails often contain sensitive personal, financial, or business data, so silent local persistence can expose users to unintended data retention, leakage to other local users/processes, or accidental inclusion in backups and logs.
