Skip to main content

What are you looking for?

Explore our services and discover how we can help you achieve your goals

Client Workspace

OWASP Top 10 in code review: a checklist for business applications

To use the OWASP Top 10 in code review, turn each of the ten risk categories into one short question that a reviewer asks only when a pull request touches that risk, and write down which answers block the merge. Netbase reviews code this way, because a ten-question checklist on every change soon becomes a box nobody reads.

Book a solution review See the related solution

By Netbase Delivery and Quality Desk · Reviewed by David (CEO) · Updated · 10 min read

star

This guide is for engineering managers, tech leads and CTOs who own code review for a business application such as an ERP, a portal, a SaaS product or an internal tool. If you want the review, scanning and testing done for you, see application security assurance.

In this guide

What does the OWASP Top 10 give a code reviewer?

The OWASP Top 10 gives a reviewer a shared list of the most critical web application risks, not a list of checks. The OWASP Top 10:2025 describes itself as an awareness document, so a team has to turn each category into a question that fits its own code before the list helps in a pull request.

The 2025 edition matters for review because the categories moved. Broken Access Control is still first and now includes server-side request forgery. Software Supply Chain Failures widens the old "vulnerable components" category to build systems and distribution, and Mishandling of Exceptional Conditions is new: error handling, failing open and logic that breaks under abnormal input.

Which question should a reviewer ask for each category?

Each prompt is one question a reviewer can answer from the diff, plus a note on where to look and what the diff alone cannot show.

OWASP Top 10:2025 category Review prompt Where to look in the diff What the diff cannot show
A01 Broken Access Control Does the server check that this user may act on this record, every time? New routes, controllers, queries that take an id, outbound URL fetches Whether the role model itself is right
A02 Security Misconfiguration Does this change a default, a header, a debug flag or an exposed endpoint? Config files, environment templates, framework settings The real production configuration
A03 Software Supply Chain Failures Is every new or upgraded package needed, pinned and from the expected source? Lock files, build scripts, CI workflow files Known vulnerabilities, which a dependency scanner finds
A04 Cryptographic Failures Is sensitive data hashed or encrypted with a current, library-provided method, with no key in code? Password handling, token generation, storage of personal data Key management and rotation outside the repository
A05 Injection Does user input reach a query, command, template or interpreter without parameters or encoding? Raw SQL, shell calls, string-built queries, templates Paths the reviewer did not trace, which static analysis follows
A06 Insecure Design Could a valid user abuse this flow, for example by skipping a step or repeating an action? New workflows, approvals, payments, limits Threats in the wider design, which a threat model covers
A07 Authentication Failures Does this touch login, sessions, password reset, tokens or MFA, and keep the existing controls? Auth middleware, session settings, reset and invite flows Brute-force behaviour, which testing shows
A08 Software or Data Integrity Failures Does the code trust data, files or updates it has not verified? Deserialisation, imports, webhooks, auto-update logic Whether signing is enforced in the pipeline
A09 Security Logging and Alerting Failures Is a security-relevant action logged without secrets or personal data, and can someone act on it? Logging calls around login, permission changes, exports Whether anyone reads or alerts on the logs
A10 Mishandling of Exceptional Conditions If this step fails halfway, does the code fail closed and leave the data consistent? Try and catch blocks, retries, partial writes, default branches Behaviour under real load and timeouts

How do you add the prompts to your reviews?

  1. Tag the risk surfaces once

    List the folders and file types that touch each category: auth code, data access, config, lock files, logging. Review tools can route those paths to a reviewer.

  2. Trigger prompts by surface, not by habit

    A pull request that changes only a report layout gets no security prompts; one that adds an endpoint with a record id gets A01, A05 and A09. Short, relevant prompts get read.

  3. Write the merge rule before the first finding

    Block on a confirmed A01, A05 or A07 issue in new or touched code, and on any secret in code. Record lower-risk findings as tickets with an owner and a target release.

  4. Hand off what review cannot settle

    Send dependency questions to a scanner, untraced input paths to static analysis, flow abuse to the threat model, and runtime behaviour to testing. The secure development lifecycle guide places those controls by CI stage.

  5. Review the prompts each quarter

    Drop prompts that never find anything; add one when an incident or a penetration test shows a gap.

What should block a merge, and what should only be noted?

A merge should stop for a confirmed issue that lets the wrong person reach data or actions, lets input run as code, or weakens authentication in code the change adds or touches. Everything else becomes a ticket with an owner, so the review stays a quality step rather than a negotiation.

The OWASP Authorization Cheat Sheet gives the access rules a reviewer can hold to: deny by default, check permissions on every request, and never rely on client-side checks. A missing server-side check on a new endpoint is a block, whatever the deadline.

For injection the rule is just as plain. The OWASP Query Parameterization Cheat Sheet names parameterized queries as the defence against SQL injection, so a string-built query with user input is a block, not a style comment.

Old issues found in untouched code are different. Log them, rank them and fix them in planned work; blocking an unrelated change for them teaches the team to avoid touching risky files.

When is a checklist review the wrong tool?

A checklist review is the wrong tool when the risk is outside the diff. A new product, a new integration with payment or personal data, or a change to the role model needs a threat model first, because the reviewer cannot see a missing control in code that was never written.

It is also the wrong tool for a large inherited codebase with no tests. Reviewing every file against ten prompts takes weeks and finds what a scanner would find in hours; start with dependency and static scans, then use the prompts on new changes. Before an audit or a customer security review, add a penetration test rather than more review.

Where AI fits

AI helps reviewers with volume. A model can summarise a large diff, point to the files that touch a tagged risk surface, suggest which prompts apply and draft a plain-language explanation of a finding and a fix. Engineers still decide whether a finding is real, set its severity and approve every change, because a confident wrong answer about access control costs more than a slow review.

Code written with AI tools needs the same prompts plus a few more. The security review gate for AI-generated code adds dependency verification and named sign-off for AI-touched changes. Netbase works with the major commercial and open-source AI tools and models, chosen per project, and runs AI-assisted engineering under human review.

What delivery record exists, and what does not?

Netbase's published security practices are secure code review and version control, TLS in transit and AES at rest, role-based access control, MFA for admin dashboards, vulnerability scanning and penetration testing, and disaster recovery; NDAs, data processing agreements and SLAs are available on request. Security by design is part of how Netbase delivers its own projects.

Netbase holds ISO 27001 certification and a SOC 2 Type II attestation for its own operations. These describe how Netbase works; they do not certify a client's application or its code review.

Delivery follows a six-step lifecycle from discovery and architecture planning to ongoing support, with weekly reviews and a named project manager, where open findings and merge rules are reviewed. Netbase also works as offshore development partner on a multi-tenant cloud ERP for a US client, the kind of business application where record-level access checks are the first review prompt.

Most Netbase projects are delivered on fixed-price contracts. No Netbase record publishes finding counts, review times, defect rates or before-and-after results, and none is offered here as proof that a review prompt works.

Alternatives: who reviews, and how to choose

There are three common set-ups for security in code review.

  • Every developer reviews with the prompts. Best for a stable team with clear risk surfaces. Risk: prompts fade when deadlines press, unless the merge rule is written down.
  • A security champion reviews tagged changes. Best when one or two engineers have more security depth; tagged paths route to them. Risk: the champion becomes a bottleneck and a single point of knowledge.
  • An outside partner reviews and fixes with the team. Best when there is no in-house security skill, or before an audit, a customer review or a major release. Ask how findings are ranked, who fixes them and how fixes are proven. See quality engineering and testing for testing that runs alongside each release.

Choose by three criteria: who knows the code well enough to tell a real finding from noise, who can keep the prompts current, and who will own the merge rule a year from now. For a fuller requirement set than ten prompts, the OWASP Application Security Verification Standard lists verifiable security requirements for web applications and services.

Limits of this guide

  • The prompts cover web and business application code, not mobile binaries, firmware or machine learning models.
  • The OWASP Top 10 changes between editions; check the current list and category names before you adopt the prompts.
  • Code review finds a share of issues; it does not replace scanning, testing or a threat model.

Plan the next step with a Netbase consultant

Frequently asked questions

No. It names the risk categories but not the checks, so a team has to turn each category into a question that fits its own code. For a complete requirement set, teams add the OWASP Application Security Verification Standard.

No. Ask only the prompts that match what the change touches, such as access control for a new endpoint or supply chain for a lock file change. A change to a report layout needs no security prompts at all.

Block a confirmed access control, injection or authentication issue in code the change adds or touches, and any secret committed to code. Record other findings as tickets with an owner and a target release.

No. Scanners find known patterns and vulnerable packages quickly, but they miss business logic, such as a user approving their own request.

Next step

If you want review prompts and merge rules set up for your codebase by the team that will fix the findings, book a solution review.

Application security assurance for web, SaaS and AI features Application security assurance for web, SaaS and AI features

Netbase provides application security testing for product and engineering leaders to find and fix vulnerabilities before release, including in AI features. We review code with AI-assisted triage, scan dependencies and test running applications, then rank findings and retest fixes. Netbase's own alignment with ISO 27001 and SOC 2 Type II covers Netbase's operations, never your application.

Learn More
line
AI-assisted QA and software testing that run inside delivery AI-assisted QA and software testing that run inside delivery

Netbase provides software testing and QA services for product and commerce teams to release often without breaking what customers use. QA runs inside delivery, not after it: acceptance criteria, automated regression with AI-generated test drafts, performance budgets and release readiness checks. Page speed shows the result: USticker's page load fell 21%, and PrintLeo's improved 35%.

Learn More
line
AI-assisted DevOps and platform engineering for routine releases AI-assisted DevOps and platform engineering for routine releases

Netbase provides DevOps and platform engineering for SaaS and commerce teams to make releases routine instead of risky. We set up CI/CD pipelines, repeatable environments and observability on AWS, Google Cloud, DigitalOcean or Cloudflare, with AI assistance for tests, reviews and incident triage, so every change ships the same way and problems surface before customers notice.

Learn More
line
SaaS product accelerator: launch an AI-ready SaaS on proven Netbase modules SaaS product accelerator: launch an AI-ready SaaS on proven Netbase modules

A SaaS product accelerator is a set of reusable Netbase modules for accounts, billing, roles and integrations that helps founders and product teams launch subscription software faster, with room for in-product AI from the first release. Reusing these modules can shorten development time, and the approach is proven on Printcart, Netbase's own web-to-print SaaS.

Learn More
line

Work with Netbase remotely

Every Netbase engagement is delivered remotely, so clients in any country and any industry can work with us; Netbase has worked with clients in 80 countries. Contracts and invoices are issued by Netbase JSC, and payment is made by bank transfer to Netbase JSC's company account or PayPal Business. Talk to the CEO, Huy Nguyen (David), on WhatsApp or by email.

Contact Netbase

Discuss a project

Netbase JSC helps organizations design, build, modernize, and operate digital products and AI-enabled business systems.
Project enquiries

project@netbasejsc.com

WhatsApp

+84 937 869 689

Office address

91 Nguyen Chi Thanh, Dong Da, Hanoi, Vietnam

Get in touch

Tell us what you want to build, modernize, or operate.

Tell us what you want to build, modernize, or operate.

Contact Netbase