Security audit
Security checks for vulnerabilities and agentic risk
This email integration’s mailbox access, automatic replies, and background connection are disclosed, configurable, and consistent with its purpose.
Choose notify or tool mode unless you want incoming mail to trigger replies. Use an inbox-scoped key for one mailbox; supplying an account key normally creates a pod-scoped credential that can access other inboxes in that pod. Configure sender and outbound rules as needed. The skill also instructs automatic command/error reporting to OpenMail, excluding correspondence and keys.
The executable command reports a bug with an endpoint, error code, and diagnostic message. It does not list threads or perform another declared email task; the reference to listing threads appears only inside the report text. Although related to OpenMail, standalone bug reporting is a materially different purpose from the described email functionality.
The description explicitly activates the skill for phrases such as "sign up" and "what did that PDF say" even without mentioning email. These requests can concern website registration or local documents, and no exclusion conditions distinguish those cases from email-related tasks.
Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.
- Use context you have. Told the user about mail from alice@example.com and
they say "reply to her"? You know the sender and thread. Just do it.
- Never ask the user for ids or addresses you can look up
(`threads list`, `threads get`).
- Reply in the existing thread. New threads only when explicitly asked.
Detected: suspicious.dangerous_exec
const child = spawn(process.execPath, [resolveBundledCliPath(), ...args], {