Nobody Is Targeting You. That's the Problem: Singapore's Incident Spike Runs Through the APIs You Forgot
On 13 November the Cyber Security Agency of Singapore put out an advisory to all organisations with an unusually blunt number in it: since September, the average number of cyber incidents reported each month has been three times the January–August average. A day earlier, CSA and the Personal Data Protection Commission had issued a joint advisory on insecure APIs, the plumbing that lets your website, app, CRM and vendors pass customer data to each other.
Read separately, they look like routine government guidance. Read together, they describe one problem. CSA says the spike does not look like a coordinated campaign against Singapore. It is opportunistic: attackers sweeping the internet for anything exposed and weak, increasingly with AI doing the sweeping. And the thing most often left exposed and weak, judging by the regulator's own case list, is an interface nobody in the business remembers is there.
"We're too small to be a target" is correct, and irrelevant
It is a common refrain among SME owners, and on its own terms it is right. Very few attackers are sitting in a room choosing your company. CSA's point is that they no longer need to. The advisory describes attackers scanning the internet for vulnerable systems, which catches organisations that were never singled out, and using AI to speed up the work: combing public code repositories for leaked credentials, writing more convincing phishing, and finding and exploiting software flaws faster and at greater scale.
That changes the economics. When discovery is close to free, being uninteresting stops protecting you. What matters is whether something you own answers when a scanner knocks. The gaps CSA lists are not exotic: unpatched or misconfigured internet-facing servers, open remote access ports, exposed APIs, missing multi-factor authentication, and weaknesses in systems run by your vendors. Some of the recent incidents, it says, came in through compromised third-party suppliers.
The breaches the regulator chose as examples
The API advisory is worth reading for its footnote alone. To illustrate the risk, the PDPC and CSA cite two local cases. At Marina Bay Sands, a misconfigured API endpoint and weak default credentials let an attacker take personal data belonging to about 665,495 people. At Tech in Asia, a legacy API that did not check authorisation tokens was used to scrape data on about 220,000 people.
Neither was a clever zero-day. The MBS case is the more instructive one. According to reporting on the PDPC's decision, the problem traced back to a large software migration in March 2023, during which an API setting was left out of the new system. The gap sat there for roughly six months. The PDPC noted that one employee had manually compiled the API configurations with no second check, called the failure negligent, and fined MBS S$315,000. A company with a large IT budget lost track of one endpoint during a routine project. Most SMEs have far less oversight than that.
The advisory's two illustrative scenarios say the same thing in a different way. In one, researchers found a public API protected only by default credentials, logged in, and pulled personal data from a database of millions of records, unnoticed, because everything they did used legitimate functions and valid credentials. In the other, internal-only endpoints became reachable from the internet when a network restriction was removed to make a third-party integration work, and nothing replaced it. The organisation found out when the incident did.
Your vendors' APIs are your attack surface too
That second scenario should worry anyone whose business runs on SaaS and integrations, which is nearly everyone. The link is usually added under deadline pressure, by a vendor or contractor, with a rule loosened "just for now". Earlier this month, Singapore Airlines said about 46,000 bookings were affected through a vulnerability at its reservations technology provider Amadeus, with several other Asian carriers caught by the same issue. SIA stressed its own systems were not breached. Its customers' itineraries were exposed all the same.
Here is where I'd push back on how most SMEs read this kind of guidance. The instinct is to buy something: a web application firewall, an API gateway, a monitoring product. The advisory does mention a WAF, but as one option inside its third measure. Its first measure is far less glamorous: “know what APIs exist and what personal data they process”, who owns each one, and which third parties can reach it, and switch off the ones you no longer use. You cannot protect, log, rate-limit or test an endpoint you don't know about. Tools layered on top of an incomplete inventory mostly protect the parts you already understood.
The other point worth absorbing is about credentials. The advisory says monitoring should cover requests made with valid credentials, because attackers misuse them. A lot of small-business security quietly assumes that a successful login means the right person. Combine that with CSA's warning about AI tools trawling public repositories for leaked keys, and an API key sitting in an old GitHub repo or a config file starts to look like a front door with the key taped to it.
What an SME should actually do this quarter
Start with an outside-in view. List every domain, subdomain, IP address and cloud service you pay for, then check what each one exposes to the internet, including the forgotten staging site and the integration a contractor set up in 2023. That is the attacker's map; it should be yours first. For each API that touches customer data, write down its owner, what it returns, and who else connects to it. The PDPC's guidance asks for exactly that inventory, and it is the first thing you would be asked for after a breach.
Then make change a security event. The MBS breach and the advisory's integration scenario both started with a routine change: a migration, a loosened firewall rule. Any time a control is removed or relaxed to get something working, someone other than the person making the change should confirm what replaces it, and test it afterwards.
Hunt for your own leaked secrets before someone else does: scan your code repositories and shared drives for API keys and passwords, rotate anything you find, and make sure each key can be revoked quickly. Put phishing-resistant MFA on admin, email, cloud and remote access, which CSA lists as an immediate measure. Turn on logging that records who called what and how much data came back, and set at least one alert for unusual volume, even from a valid account.
Finally, ask your key vendors two questions in writing: how will you tell us if your systems expose our customers' data, and how fast? If cost is the blocker, CSA's advisory points to its CISO-as-a-Service programme, which co-funds up to 70% for eligible SMEs. None of this is a product purchase. It is knowing what you have, and noticing when it changes.
Belian Advisory runs external attack surface discovery for SMEs, mapping what your domains, APIs and vendor integrations actually expose before a scanner does; if you're not sure what answers when someone knocks, get in touch.