API security testing checks that every endpoint confirms who is calling and whether that record is theirs, including the old and forgotten ones. Software supply chain security covers what goes into your builds: secrets, open-source components and the pipeline itself. We test under written authorisation, help your engineers fix the patterns behind each finding, and add checks to CI/CD so gaps do not come back.
Key takeaways
- The leading API risk in OWASP's 2023 list is an endpoint that hands over any object ID it is given, whoever is asking.
- Old versions, partner integrations and test servers are part of your attack surface until they are retired.
- A leaked key has to be revoked and reissued; removing it from the code leaves every existing copy working.
- An SBOM tells you within minutes whether a newly disclosed vulnerability affects what you ship.
- Checks in the pipeline keep fixes in place; a one-time test does not.
A modern product is mostly plumbing you cannot see: dozens of APIs behind the app, hundreds of open-source packages inside each service, and a build pipeline that turns a commit into a release in minutes. Attackers know this. Rather than break encryption, they ask an API for record 1002 after being given 1001, find a cloud key pasted into a repository, or wait for a package you depend on to be hijacked. Nothing breaks on screen, so these gaps can sit open for a long time.
We test those layers, help your engineers fix the cause rather than the symptom, and wire the checks into the tools your team already uses.
Where the real risks sit
The 2023 OWASP API Security Top 10 ranks broken object level authorisation as the leading API risk: a service hands back whatever object ID it is given without asking whether this caller should have it. Around it sits a familiar cluster of problems:
- Trusting the ID. Swap one invoice or report number for another and the API returns it.
- Sending everything. The response carries every column and relies on the front end not to display some.
- Hidden admin routes. Staff-only operations that any signed-in account can call.
- Zombie endpoints. Version one of the API, a partner feed or a staging host, still live and still connected to real data.
- Credentials in the open. Tokens and passwords in source files, tickets, chat threads or CI output.
- Weak links upstream. Stale packages with published flaws, unverified container bases, and pipelines that hold far more access than they need.
The price is paid in incident work, customer letters, stalled enterprise deals and sprints lost to emergency fixes.
How we run it
We test only what you own or are entitled to have tested, and only after a signed authorisation that fixes the scope, the testing windows and who to call. Third-party APIs are handled through the provider's own programme. Inside that boundary the work has five steps.
Every API, including forgotten ones
Check who, and whose record, every time
A vault, rotation, no keys in code
SBOM and vulnerability scanning
Checks in CI/CD, signed builds
- Discover. We list every API that answers, not just the documented ones, by comparing specifications and repositories with what the gateways, DNS zones, cloud accounts and your own apps' traffic reveal. Each one gets an owner, a data label and a decision: keep, fix or switch off.
- Authenticate. With two or more test users per role, we try to fetch and edit each other's objects, reach operations above our role and pull fields we should not get. The repair goes into a single authorisation layer that every route passes through.
- Protect secrets. We search code, commit history, CI output, images and config for live credentials. Anything found is revoked and reissued, because deleting a line does not delete the copies. Going forward, services draw short-lived credentials from a vault, one identity per service.
- Know your components. Each build produces a software bill of materials. It feeds a vulnerability scanner, and findings are triaged by severity, assigned and tracked to closure.
- Secure the pipeline. Secret, package and image scans and the cross-user authorisation tests run on every merge. A release with critical findings is blocked, pipeline tokens are narrowed, and artefacts are signed so any deployment can be traced back to reviewed source.
SBOMs and the software you buy
Your exposure also includes software from suppliers. In its SBOM guidelines, version 2.0 of July 2025, CERT-In advises government bodies, public sector organisations and essential services to make SBOMs part of how they buy software. If that advice is not addressed to you, it is still worth asking important suppliers for an SBOM and their patching process. And when your own customers ask you the same question, your SBOM is the answer.
AI agents add another caller, since they hit your APIs on a user's behalf; AI red teaming tests that route. Plant equipment and connected devices fall under OT and IoT security, and recovery after a breach under cyber resilience.
What you get
| Deliverable | Contents | Use |
|---|---|---|
| Live API register | All reachable APIs and versions with owner, data label and keep, fix or switch-off decision | Defines what must be protected, and stays current |
| Access control findings | Cross-user and cross-role issues per route, rated, with steps to reproduce and a recommended repair | Feeds the engineering backlog |
| Credential clean-up record | Secrets found, revoked and reissued, plus the vault set-up and its access rules | Shuts the leaked-key door |
| Component inventory | An SBOM for each service and image, linked to open vulnerabilities with owners and due dates | Lets you answer "are we affected?" in minutes |
| Pipeline guardrails | Merge-time scans, release blocks, narrowed pipeline tokens and artefact signing | Stops fixed problems from returning |
| Supplier checklist | Questions and contract points on APIs, components and patching for your vendors | Supports procurement and vendor reviews |
How to start
Pick the customer-facing APIs that handle your most sensitive data, together with the pipeline that ships them. Share the specifications, gateway settings and repository access, and we will agree a signed scope and test accounts. To set that up, get in touch.
The DPDP side of this
Our DPDP practice, DPDP Ready, has a detailed guide on how this work maps to India's data protection law. Read it on DPDP Ready →
Frequently asked questions
What is the difference between API penetration testing and an automated API scan?
An automated scan sends known bad inputs and flags common misconfigurations quickly and cheaply. A penetration test adds a person who understands the business logic, uses several test accounts and tries to read or change records that belong to someone else. Authorisation flaws are mostly found that way, so a sound programme uses scans continuously and manual tests on the APIs that matter most.
How do you find shadow APIs?
Compare what is documented with what is actually running. Useful sources are gateway routes, DNS zones, cloud account listings, the calls your own apps make, code and partner connection settings. Anything that answers on the network but is missing from the inventory is a shadow API to be owned, secured or retired.
What is a software bill of materials used for?
An SBOM lists the components inside an application or container image, with versions and suppliers. Its main day-to-day use is speed: when a vulnerability is disclosed in a library, you can see at once which products contain it and prioritise the fix. It also answers customer and procurement questions about what you ship.
How do we stop secrets being committed to Git?
Use a pre-commit hook and a server-side scan that block commits containing keys or passwords, and scan existing history too. Move every secret into a vault and have services fetch short-lived credentials at run time. If a secret does slip through, rotate it immediately and treat the old value as compromised.
How often should APIs be security tested?
Automated checks should run on every build. A deeper manual test makes sense before a major release, after significant changes to authentication or authorisation, and at a regular interval for APIs that carry sensitive data. New or changed endpoints should never ship without the authorisation tests in the pipeline.
Scope it with us
A 30-minute call to understand your systems and what good looks like. We aim to send a written, fixed-scope proposal within three working days.
Related services
This page describes engineering and security services. Where it mentions laws, rules or standards, it is general information, not legal advice.