Editorial and evidence policy

Global evidence, not one-country assumptions

ScamMark is written for an international audience. General guidance should describe principles that travel well across borders. Country-specific legal, reporting, banking or dispute procedures must be clearly labelled and must not be presented as universal.

Primary-source preference

ScamMark prefers official sources for reporting procedures, enforcement actions, regulatory status and platform-specific instructions. Source links are included on guides where they materially support the advice.

Research beyond English

ScamMark reviews useful research and official material regardless of publication language when it can be interpreted reliably. Non-English sources may reveal local scam patterns, terminology, enforcement data or prevention approaches that English-only research would miss.

Where a translated term could change the meaning, ScamMark should preserve the original concept, avoid overstating the translation and link to the original source.

Research interpretation

Academic papers, surveys and enforcement statistics are evidence, not decoration. ScamMark should identify important limits such as geography, sample population, self-report bias, experimental design and whether a result establishes association rather than causation.

A finding from one age group, country or platform should not silently become a claim about everybody.

Data, charts and images

ScamMark may create its own charts from published data when doing so improves understanding. The original dataset, paper or report must be cited, and any transformation or calculation should be understandable from the surrounding text.

ScamMark does not copy third-party charts, photographs or illustrations simply to make a page look complete. Original diagrams, illustrations and site imagery should use a consistent visual system and should clarify the subject rather than sensationalize it.

Automated checker evidence model

ScamMark checker results must separate verified facts, observed technical signals, deterministic rules, heuristic warning signals, unavailable or unverified information, and limitations or absence of evidence. A finding should explain both what was observed and what the observation does not prove.

Attention labels such as Informational, Worth checking and Potential concern describe how urgently a signal deserves verification; they are not legal findings or declarations that a person, business, message or website is fraudulent.

Deterministic rules are reserved for defined observable patterns. For example, a fake-verification flow that instructs a user to open a system shell and paste or execute commands can trigger the ScamMark ClickFix Safety Rule. Heuristics such as urgency language or unusual payment methods require more context and must not be presented as proof.

A lack of triggered rules is absence of detected evidence, not evidence of safety. Website retrieval can also be incomplete or differ from what another visitor sees. ScamMark does not convert a clean automated fetch into a claim that a site or message is legitimate.

Pasted-message analysis should remain browser-side whenever the current rules can operate locally. Checker measurement must not include the submitted message content or extracted URLs.

Scam messages, screenshots and real-world specimens

A real message, email, advertisement, invoice or conversation can be useful evidence of a scam pattern, but publication requires context. Personal data should be redacted, active malicious links should be defanged or handled safely, and materially identical variants should be clustered rather than published as thin duplicate pages.

A specimen can show what was sent; it does not by itself prove the legal identity of the sender. ScamMark separates the observable message pattern from claims about who was responsible.

Allegations and user reports

A complaint, post or user report can be useful as a lead, but it is not automatically proof. ScamMark distinguishes verified findings from allegations and unverified reports.

Corrections and review dates

Reporting contacts and platform procedures can change. Pages with jurisdiction-specific or provider-specific instructions should show a review date and be updated when authoritative information changes. Material corrections should be made in the source and carried through the normal build and validation process.

No guaranteed recovery

ScamMark does not promise to recover funds and does not endorse anyone merely because they claim to trace or recover cryptocurrency, transfers or other losses. People who have already lost money are frequent targets of refund and recovery scams.

No paid verdicts

A commercial relationship must never buy a favorable safety verdict, suppress a legitimate warning or determine editorial conclusions.