Impersonation Detection Console

brand-impersonation detection → a signed receipt for every decision
1Sensors → 2Filters + classify → 3Triage + analyst decision → 4Signed receipt every confirm, reject, escalation and release is signed the same way
Sensors: default offline fixtures. Social-network ingestion is a fixture sensor; enable “Broad live sensors” for live App Store + look-alike domains + CT.

Analyst queue (the model ranks · a person decides)

Press “Run pipeline” to detect impersonations.

When you are not sure

Loading the governance settings…

Signed evidence receipt

Run the pipeline to produce signed receipts.

OSINT reverse-pivot (keyless)

Seed a suspect domain → sibling infra + campaign signals (crt.sh · dns · urlscan).

Label ledger · threshold calibration

Analyst confirms and rejects go into an HMAC-chained label ledger. Thresholds are recalibrated from those labels, and every threshold set is versioned. The operating point marks which findings clear the calibrated threshold; it never changes a signed decision.
Confirm or reject a candidate to grow the label ledger.