Check a suspicious message before you act
SuspectFraud helps you spot warning signs, preserve evidence and report suspected scams through verified official channels.
Share a message, screenshot or link in a compatible assistant conversation and ask, “Does this look like a scam?” The quick check explains visible warning signs and suggests one practical next step. It can then help prepare reports to domain registrars, relevant providers, the impersonated company and, for Singapore cases, ScamShield.
What it checks
Requests for deposits to unlock earnings, one-time codes, account credentials, remote access, pressure to act and unexplained company-domain mismatches. A warning explains the observed signal. It does not establish guilt or authenticate a sender.
You choose what to share and report
SuspectFraud is a set of assistant instructions. It has no standalone inbox, message listener or background monitoring service. Dot and other hosts can use the skill when they support it and make the relevant content available. Installing it does not guarantee every message will be checked.
Reports require your authority and access to the destination. A prepared report is shown separately from a confirmed submission. The assistant should retain receipts and tell you when a login or device action is still needed.
Keep sensitive evidence private
Remove passwords, one-time codes, recovery phrases and unrelated personal details before sharing. Use redacted copies for reports and keep originals privately. Do not publish your evidence in a GitHub issue.
Availability
The source is public on GitHub. The ChatGPT plugin is currently private to its publisher's account; a public directory listing is being prepared. The skill does not require a separate SuspectFraud subscription. Your assistant host may have its own charges and access requirements.