For Singapore SMEs · Security, evidenced

Find out what your security already proves.

Most security questionnaires start with “which certification do you want?” and end an hour later with a folder of unanswered questions. We start with the other end: the systems you run, the records you already keep, and the people who can answer for them.

  • No target required to start
  • No administrator account needed
  • Nothing is read before you agree

No account, no agent, no install. You will leave the first call with a written picture of what you know, what you do not, and who owns each gap.

Two colleagues reviewing a document together at a desk in a small office Evidence First Security Passport
Small teams start with a conversation and a handful of dated records — not an administrator handover.

No target required

A useful, honest picture of your own security — before you pick any framework, if you ever do.

No jargon at the point of choice

Every objective, permission and gap carries one plain sentence: what we read, why it matters.

No silent connection

Nothing is read from a provider until you have seen the exact list, previewed it and agreed.

Nothing is inferred

A missing item reads “unknown”, never “pass”. Unknown gets an owner and a date, like any other gap.

The problem we are solving

“Which certification should we get?” is the wrong first question.

It puts the standard in front of the organisation instead of the organisation’s own obligations, services and evidence. Three things usually go wrong.

Problem 01

The questionnaire marathon

A 200-row form lands on a business owner who does not know what a hardened endpoint is. Most cells get guessed, and a guess reads like an answer.

Problem 02

The consultant spreadsheet

Evidence is re-keyed by hand every cycle, with no source, no date and no owner — so the same gap is paid for again in six months.

Problem 03

The dashboard that scores what it cannot see

A green percentage over a denominator nobody agreed. Missing evidence, denied access and stale records quietly become a pass.

Security Passport inverts the order. Start with what you do, what you run and what you already keep. The framework comes later, as a suggestion you may accept, reject or ignore. Only the certification body can issue a mark — we never can.

How it works

Five steps. The first ten minutes needs nothing but you.

Step one is a conversation, not a form. You can stop after step two and still have something useful.

  1. Talk — four plain questions

    What does your organisation do, roughly how many people and devices are involved, which system would hurt most if it stopped for a day, and who can decide about security spending. Every question accepts “not sure”.

    You do
    One 15–20 minute call, as a sponsor rather than a technician.
    We do
    Write down your answers as scope, obligations and named owners.
    You get
    A written first picture, even if every technical answer is still unknown.
  2. Posture picture — before any target exists

    The dashboard opens on what is known, what is unknown and where the biggest gaps sit: coverage by service, site and tenant, plus the highest material gaps. No objective is selected and none is required.

    You do
    Correct anything we have written down wrongly.
    We do
    Separate known, unknown, pending and confirmed out of scope.
    You get
    A baseline you can compare every later month against.
  3. Choose zero, one or several objectives

    We suggest — Cyber Essentials, a Cyber Trust tier, NIST CSF, ISO 27001, CIS IG1 or none of them. Each card explains why it fits, what it covers, what evidence it costs and what stays unresolved. Selecting nothing is a legitimate outcome.

    You do
    Pick, defer or decline. Nothing is auto-enrolled.
    We do
    Show the trade-off: scope, effort, evidence burden, alternatives.
    You get
    A decision you can defend to a customer, an auditor or your own board.
  4. Bring evidence the easy way

    Upload a file, hand the task to your IT provider, or approve a guided read-only connection — per source, in that order of convenience. You can use one route for everything or mix them freely.

    You do
    Pick the easiest route for each source, and skip what you do not have.
    We do
    Preview the rows, flag secrets and stale dates before anything is admitted.
    You get
    Evidence with a source, a date and a scope — not a row in a spreadsheet.
  5. Review, assign, recheck

    Each gap gets one owner, one date and one next action. Fresh evidence closes it once; a confirmed regression reopens it. Expiry means stale, never failed.

    You do
    Work the top three actions, not the whole list.
    We do
    Track what changed, when it changed and who accepted it.
    You get
    A record that survives a staff change, an audit or a customer audit.

Evidence access

Three routes to the same record. All three are first-class.

A connector is a shortcut, never the only path. Choosing one route never obliges you to use another.

Hand-off

Your MSP, IT provider or auditor does the work

A written request with the exact fields, scope and due date. We write it; they answer it. The result comes back as a record with a source and a date.

Good for: items only a technical role can produce, or anything you would rather not export yourself.

Guided read-only connection

Your IT, once, with your approval

An itemised, read-only permission list you approve before anything is read. The first read returns a preview you can reject, narrow or redact. One-time and recurring are separate choices.

Good for: repetitive or wide sources, and period data only the provider holds. Revoke any time — future reads stop and stored tokens clear.

Before any upload lands. The form states in one line what must not appear: passwords, API keys, private or recovery keys, tokens, connection strings, Wi-Fi passphrases, session cookies, mailbox or file contents, customer records and full device configurations. If a secret is found in a file that already arrived, it is quarantined, treated as exposed, and you are asked to rotate it — we never quietly strip it and move on.

The rules we hold ourselves to

Six things Security Passport will never do.

These are acceptance conditions, not marketing lines. Each one has a check behind it.

Require a target

Zero objectives still gives you a usable, honest dashboard with real gaps on it.

Require an administrator

The first evidence route is a file a non-technical person can produce on their own.

Connect silently

No provider API call happens before an explicit agreement exists for that read.

Assume continuity

Every provider route is optional, separable and removable. You can revoke end to end.

Delete quietly

Revoking stops future reads and clears stored tokens; accepted evidence stays visible with its source and date.

Infer a pass

A missing item reads “unknown”. Denied, partial and stale access are never counted as compliant.

Two starting points

Where are you today?

Most organisations are in one of these two places. Both routes use the same five steps — only the starting evidence and the objective look different.

Journey A · no IT team

Eleven people, no administrator, one lost laptop

You run the business day to day. Nobody owns security and nobody should have to. This journey starts with a 20-minute conversation and a screenshot, and ends with a written picture of where you stand against Cyber Essentials — without a single agent installed on a staff machine.

  • What to collect in week one, and exactly where each item lives on screen
  • What to ask an IT provider to produce for you, in writing
  • Which questions the certification actually asks — and which ones it does not

See the no-IT-team journey

Journey B · already certified

Essentials passed, now a customer wants Trust

You already hold Cyber Essentials and a large customer has asked for Cyber Trust. The gap is not knowledge — it is evidence over time: operating records, logs, supplier artefacts and a scope decision. This journey lays out a 90-day evidence plan and where Promoter actually differs.

  • What the tier ladder really adds, statement by statement
  • The three things that most often stall a Trust application
  • How to reuse your Essentials evidence without pretending it is equivalent

See the Essentials-to-Trust journey

A worked example

Selegie Roastworks, eleven staff, no IT department.

A composite example built from the kind of first conversations we run. It is illustrative, not a real customer.

Selegie Roastworks Pte Ltd

Specialty coffee roaster · Jurong · 11 staff · founder-led

People
11
Devices
14 known, 3 unknown
Tenants
2 (Google Workspace, Shopify)
Chosen objective
None — yet

“We knew we had to do something about MFA. What we did not know was that half our staff were already signing in with their phones, and nobody could tell us which half.”

The founder answered the four plain questions on a Tuesday morning call. Nine minutes later, the first picture existed: a booking system nobody could name an owner for, a shared password in a spreadsheet, three laptops that had not been seen since March, and no idea whether the shop-floor tablet was encrypted.

Nothing was connected and nothing was asked of an administrator. Selegie uploaded two dated screenshots — the Google Workspace security settings page and the Protection-history screen on one office laptop — and answered “not sure” to everything else. Those answers became five evidence requests, each with an owner and a date, rather than five blank rows.

Six weeks later the same founder could say: the Shopify admin account is protected by a passkey, the shared spreadsheet is gone, the three unseen laptops are written off and replaced with two managed machines, and the shop tablet is still unknown — because nobody had been asked yet.

The most useful line on the dashboard was never a percentage. It was: “1 device unknown — owner: your IT provider — asked 14 May — not a failure, a question waiting for an answer.”

Frameworks we map against

Reference, not a badge.

We track requirements against the objective you choose, and we keep the source version and clause identifiers with each record. Mapping never gives readiness credit on its own.

CSA Singapore

Cyber Essentials

CSA Singapore · baseline · general scheme

The usual first suggestion for a Singapore SME: a small, concrete set of controls, independently verified by desktop assessment, valid for two years.

CSA Singapore

Cyber Trust

CSA Singapore · five tiers · two-stage audit

Supporter, Practitioner, Promoter, Performer, Advocate. Evidence is gathered over time and reviewed by a certification body — a different job from passing a self-assessment.

NIST

NIST CSF 2.0

Outcome framework · not a certification

Govern, Identify, Protect, Detect, Respond, Recover — useful for board reporting and for prioritising work where no formal certification is needed.

ISO / IEC

ISO/IEC 27001:2022

ISMS certification · scope matters

Worth it when procurement demands independently certified governance. Scope and effort decide whether it suits you — an Annex A checklist on its own does not.

Singapore Government

Singapore Government ICT&SS

Control catalog · 17 groups

Relevant if you deliver in-scope government digital services. The contract and the agency owner decide what applies — we track it rather than guess a scope for you.

Framework names and links point to the issuing bodies. Security Passport is not accredited by, endorsed by, or affiliated with any of them.

Straight answers

Questions we get asked first.

Will this get me certified?

No, and nothing we produce can. Certification is issued by the certification body after its own assessment. What we do is make the evidence for that assessment obvious, dated, owned and complete — so the assessment is a formality rather than an excavation.

Do we have to choose a framework before we start?

No. The first conversation and the first posture picture work with no target at all. When you do pick one, it is recorded as a suggestion you accepted — never auto-selected, and you can pursue several or none.

Do we need an IT administrator, an agent or a new licence?

No, no and no. The first evidence route is a file or a dated screenshot produced by a non-technical person. Connections to Microsoft 365, Google Workspace or AWS are optional per source, read-only, and revocable — and they never require a second licence to be useful.

What will you never ask us for?

Administrator passwords, private or recovery keys, API tokens, connection strings, Wi-Fi passphrases, session cookies, mailbox or file contents, customer records and full device configurations. We never read a provider until you have seen an itemised permission list and agreed to it.

What happens to the evidence if we cancel?

Revoking a connection stops future reads and clears the stored connector credentials, and we show you how to revoke the provider grant yourself. Evidence you already accepted stays visible with its source and date, because a record that vanishes is worse than a record that is wrong. Retention and deletion of imported records are explained separately, before you upload.

Is my data used to train a model?

Not by default. Any AI-assisted explanation is scoped to the tenant, the actor and the purpose, with model provider, region, retention and fallback agreed per tenant before any customer data is used. Personal data, raw logs and raw artefacts stay out by default, and feedback becomes review data rather than an automatic rule change.

Where is our data held?

The intended deployment is in Singapore, with the application, worker and evidence stores kept off the public internet and only the ingress exposed. Specific hosting, residency and retention decisions are confirmed with you before any real data is admitted.

Start with a conversation, not a purchase.

Twenty minutes with a sponsor and no technical role required. You will leave with a written picture of what you know, what you do not, and who owns each gap — whether or not you ever work with us again.