Application security
Application security for software your team builds
Reviews, workshops and architecture work are also available for applications and APIs without agents. I start where design and code decide security, and I propose a concrete change for every finding.
- Scope
- One application or API that your team builds and runs, with sign-in, permissions, tenants, secrets and the chain from repository to deployment. We agree on the exact scope in writing during the intro call.
- Format
- As a review, as a workshop with your team or as architecture work during the build. We decide which format fits in the intro call. You then receive a proposal.
- Outcome
- Prioritized findings, each with a concrete proposal for the change. Architecture work also produces the design, and on request the code, together with your team.
Who this is for
Your team builds an application or API itself, often a SaaS product with several tenants. A large customer asks about your architecture in a security questionnaire, your first enterprise customer is coming up, or an app built with Lovable, Bolt or Cursor is about to launch.
What I do not take on
- Standalone penetration tests. I read architecture and code and see where a test should start. I do not offer a penetration test as a service of its own, though. If your application needs one, I describe the scope a testing firm should cover.
- Network and infrastructure. Where your application meets the infrastructure, for example secrets in deployment, I look at it too. Networks, firewalls and servers are better reviewed and hardened by a team that specializes in them.
- ISMS and monitoring in operations. You can use the threat model and the report as material for your evidence. I do not build an ISMS or run a SOC, and the report is an expert assessment, not a seal of approval.
- Migrations and vendor selection. I show where and how strongly you depend on a vendor. The migration itself and choosing a new vendor are not part of the work.
- Your system has an agent. Then the go-live review fits better. It targets the risks of agents, RAG and MCP.
Contents
What is included
-
Security architecture and threat modeling
Together we record where data enters your application, which components trust each other and where a boundary is missing. This becomes a threat model for the application and its APIs.
-
Authentication and authorization
I check whether your API verifies who is asking or trusts the frontend, and how OAuth and OIDC are wired in. For SaaS this includes whether the tenant ID comes from the token or from the request.
-
Tenant isolation
I look at whether every query filters by tenant, including background jobs and exports, and whether an isolation test fails as soon as one tenant sees another tenant’s data.
-
Secrets
For every key I clarify where it is stored, who can read and rotate it and whether it can show up in the frontend, in logs or in CI.
-
Supply chain, build and CI/CD
I check which dependencies your pipeline ships, how security updates reach images and lockfiles and which build steps have access to secrets.
-
Secure code review
I read the code on the paths that security depends on, such as sign-in, permission checks and data access. This is not a full pass through the entire codebase.
-
Vendor dependency and exit readiness
With your team I build an inventory of dependencies for identity, data, keys and models, and check whether backups and keys are stored outside the account they are meant to protect.
Process
How it works
-
Intro call and scope
We clarify which application it is, what prompts the work and which format fits. The proposal then sets out scope and dates in writing.
-
Reading
I read the architecture, code, configuration and the pipeline up to deployment. Read access to the repositories is enough for this.
-
Sessions with the team
With the engineers who built the application I go through data flows, permissions and secrets. That is where the questions get answered that the code alone cannot answer.
-
Report and changes
You get the findings sorted by impact, each with a concrete proposal for the change. If you like, I then work on the architecture, on my own or together with your team.
Writeups
Further reading
In these articles I describe patterns that this page is about, each with code or configuration.
Contact
Talk about your application
In the intro call we clarify which format fits your application and what scope makes sense. You then receive a proposal.