Method
We do not score portals. We agree a boundary, walk the journeys customers already use, write evidence against that boundary, and sit with you until the register is understood. This page is the method behind every engagement listed under work.
A call, usually thirty to forty minutes. We name the URL, the roles, the environment, and the six journeys that matter this quarter. If production is off limits, that limitation is written into the letter before anyone logs in. If you cannot grant test accounts, we stop — a review of screenshots you chose is not an audit.
Reviewers complete the agreed chores: sign-in, reset, account home, a bill, a document, a ticket, an appointment, a profile change. We pause where the interface, the copy, or the record disagrees. We do not ‘break in’. We use the same doors your customers use.
When a label and a field have drifted — invoice date, account number, a PDF that knows more than the HTML — we put both names in the register. You may send a payload sample or an export. We read it against the journey, not against an architecture poster.
A sponsor letter, a finding register with evidence, and a retest list. Severity follows customer harm and operational cost. We will not invent a score. Findings outside the boundary are labelled as such so Monday morning stays usable.
We brief in Kuching or by video. Owners are named in the room, not in a tool we do not share. Within six months you can commission a verification pass on the same journeys. New scope is a new audit.
This method is not a penetration test, not legal advice, and not a build contract. If a finding needs a lab, counsel, or a vendor you cannot change this quarter, the pack says so in one sentence.
If the method fits the portal you already run, start with the engagements — or send the URL and the chore that keeps landing on the counter.