Continuous monitoring of the human factor

The quick scan is a photograph of the organisation. The CPF engine follows people over time: every five minutes it reads what the security systems already record and scores the 100 indicators.

What it reads

People are identified only by a pseudonymous hash. 79 of the 100 indicators can be read from security data; the others need the assessment.

How it scores

A first layer computes each indicator from the snapshot with explicit rules, so every score can be traced back to the fields that raised it. A second layer, a small language model, reads the same snapshot as text and gives its own reading of intent (benign, suspicious, malicious) as a check on the first.

Three examples, scored live

These three snapshots are sent to the engine's API when this page is generated; the numbers below are its answers.

An administrator at work, office hours, ordinary commands on a low-value workstation.

CPF score: 1/100 | Alert: GREEN | Second reading: benign

Yellow indicators: 2.9 Shift Change Exploitation Windows

A privileged user during a round of layoffs, with a pending review, touching confidential data on a critical system.

CPF score: 20/100 | Alert: YELLOW | Second reading: suspicious

Red indicators: 3.9 Social Identity Threats Susceptibility

Yellow indicators: 1.1 Unquestioning Compliance with Apparent Authority, 1.5 Fear-Based Compliance Without Verification, 1.8 Executive Exception Normalization, 2.2 Time Pressure Cognitive Degradation, 3.2 Commitment Escalation Traps, 4.7 Anxiety-Triggered Mistakes

Night-time access, a recent breach in the company, secret data and 400 MB leaving through a web service.

CPF score: 39/100 | Alert: RED | Second reading: malicious

Red indicators: 1.1 Unquestioning Compliance with Apparent Authority, 1.8 Executive Exception Normalization, 1.10 Crisis Authority Escalation, 2.1 Urgency-Induced Security Bypass, 2.7 Time-of-Day Vulnerability Windows, 3.10 Reputation Management Conflicts

Yellow indicators: 3.5 Scarcity-Driven Decisions, 4.10 Emotional Contagion Effects, 6.3 Diffusion of Responsibility, 7.10 Recovery Period Vulnerabilities, 8.8 Archetypal Activation Triggers, 10.6 Gray Rhino Denial

Use it

The engine runs on Silicon Psyche Labs. Send one snapshot per user and time window to the API; with an account the analyses are stored and the dashboard shows each person's trend and the organisation view.

curl -X POST https://splabs.io/api/v2/cpf/analyze \
  -H 'Content-Type: application/json' \
  -H 'Authorization: Bearer psa_...' \
  -d '{"snapshot": {"snapshot": {"id": "...", "ts": "...", "window_sec": 300, "who": {...}}}}'

splabs.io Start with the quick scan