VeriTrustLab does not publish an accuracy, precision, recall, or F1 claim until a controlled, reproducible benchmark is completed. Runtime risk scores are investigation signals, not measured product accuracy.

No VeriTrustLab benchmark is currently published.

The repository does not contain a controlled labeled evaluation, confusion matrix, or verified accuracy, precision, recall, or F1 result. Accordingly, the site makes no numerical performance claim.

Current scan output is an AI-assisted triage signal. A provider model's confidence value is not a VeriTrustLab accuracy measurement and is not a legal conclusion, independent forensic certification, cybersecurity certification, or final proof.

Published model cards

MailGuard email-threat evaluation

The current SIH production path uses VeriTrustLab MailGuard as the only email-text classifier and VeriTrustLab Swift as the only URL classifier. Historical or experimental model registry rows are not part of the executable SIH product contract.

• A valid evaluation would report representative email-message slices separately and evaluate URL classification as its own task.
• It would assess model-only predictions separately from the combined rule score.
• Any benchmark must state that a low-risk result does not prove a message is safe.

Swift URL-intelligence evaluation

The available link path uses a server-configured VeriTrustLab Swift classifier plus deterministic URL-string indicators. If hosted inference fails, the current path surfaces a model failure/degraded state rather than silently substituting a local verdict. It does not fetch pages or follow redirects.

• No VeriTrustLab link-classification benchmark is currently published.
• Classifier output and deterministic URL indicators would need separate evaluation.
• Provider-failure frequency and rule-only diagnostic signals must be measured separately from successful hosted-classifier results.

Metrics required before a performance claim

• Accuracy, precision, recall, and F1 score on documented labeled test sets.
• False positive and false negative rates by scan type.
• Confusion matrices for each evaluated model and dataset split.
• Latency and provider-failure/degraded-result frequency under production-like conditions.

Dataset notes

Any future benchmark would need to identify its dataset source, labeling process, sampling approach, train/test separation, known bias, exclusions, preprocessing, and evaluation date. Without those details, a headline accuracy figure would not be reliable.

Limitations

• False positives can incorrectly flag benign content as suspicious.
• False negatives can miss risky content, especially when attackers change wording, domains, or URL structure.
• Provider failure or model-contract changes can change which evidence is available and therefore change the final score distribution.
• High-impact cases should include manual review and trusted source verification.