Skip to main content

Forensic Dashboard — the anti-hallucination layer

The Forensic Dashboard watches what a machine asserts and scores how much it can be trusted. It runs twenty pulse-coupled forensic operators — HF1 through HF20 — that score every claim your machine emits against the 1.287 Hz network clock, and rolls them into a single composite.

The composite truth score​

The dashboard's headline is S_forensic, a weighted aggregate of all twenty operators. Not every signal counts equally — the composite gives weighted priority to the ones that matter most:

HF5 — legal relevance

How much a claim carries legal weight, prioritised in the score.

HF16 — ethical severity

The ethical stakes of a claim, weighted into the composite.

HF20 — binomial truth

The core truth signal — a claim intended to bind must clear a high HF20 threshold.

The operators are pulse-coupled to the network clock, so scoring happens on the same 1.287 Hz heartbeat as everything else, and each claim's score is a clock-stamped, recorded quantity rather than a one-off opinion. High-stakes claims are held to explicit thresholds before they're allowed to bind.

Where it's used​

The same HF forensic spectrum underneath this dashboard powers other surfaces: it's what Phish-Guard maps phishing signals onto, and what Zeq Antivirus uses to score threats without a signature database. The Forensic Dashboard is the direct view of that layer — a live readout of how well a machine's outputs hold up against twenty independent measures of truth.

Open it​

Open the Forensic Dashboard →

Watch the HF1–HF20 scores and the composite, live.
  • Phish-Guard — the same forensic operators, applied to phishing.
  • Zeq Antivirus — threat scoring on the HF spectrum.
  • Operators — the HF forensic family in the catalogue.