Explore our new Cybersecurity Terms Glossary, explained by our experts.

View the Glossary View the Glossary
PRIVIA

Cybersecurity Terms Glossary

What Is OWASP ZAP?

ZAP, the Zed Attack Proxy, is a free web application security scanner. It works as an intercepting proxy, a passive and active scanner, and an automation component for continuous integration.

Illustration for the cybersecurity term OWASP ZAP
Home / Blog / Cybersecurity Terms / What Is OWASP ZAP?
17 November 2026

What Is OWASP ZAP?

What Is OWASP ZAP?

ZAP — the Zed Attack Proxy — is a free, open-source web application security scanner. Simon Bennetts started it in 2010 as a fork of Paros Proxy, it became one of the best known OWASP projects, and it remains the default answer when somebody needs dynamic application security testing without a budget line.

At its core it is an intercepting proxy. Traffic between a browser and the target passes through ZAP, which records every request and response, lets you modify them before they are sent, and analyses what it sees.

Passive and Active Scanning

The distinction governs when it is safe to run ZAP at all.

Passive scanning only observes traffic you generate anyway. It finds missing security headers, cookies without the Secure or HttpOnly attributes, information disclosure in error pages and comments, weak cache directives, and content type problems. It sends nothing extra, so it is safe against any environment and costs nothing but memory.

Active scanning attacks. It replays requests with injected payloads to test for SQL injection, cross-site scripting, path traversal, command injection, and much else. It creates records, triggers emails, fills logs, and can take an application down. Never point an active scan at production without written authorisation and a plan, and never at a system you do not own — in the UK, unauthorised testing is an offence under the Computer Misuse Act 1990, and the owner’s permission is the only thing separating a scan from a crime.

Finding the Application First

A scanner can only test what it has seen. ZAP’s traditional spider follows links in returned HTML, which works poorly on a single-page application where the routes exist only after JavaScript runs. The AJAX spider drives a real browser to reach those states, and is slower but far more complete.

The most reliable approach for a modern application is neither: proxy your existing functional or end-to-end test suite through ZAP. The tests already know how to log in, navigate, and exercise the interesting workflows, and ZAP passively analyses everything they touch and can then actively scan the discovered endpoints. If you have an OpenAPI, SOAP, or GraphQL definition, importing it directly is better still.

Authentication Is the Hard Part

Most disappointing scans are disappointing because ZAP was never logged in. The symptom is a report full of findings on the login page and nothing beyond it. ZAP supports form, script-based, JSON, and header-based authentication, plus session management and a verification request so it can detect when it has been logged out and re-authenticate. Budget real time for this configuration, and always confirm from the site tree that authenticated pages were actually reached before trusting a clean result.

Running It in a Pipeline

The Automation Framework describes a whole run as a YAML plan — contexts, authentication, spidering, active scan policies, thresholds, and report generation — which can be committed alongside the application. Simpler is the packaged baseline scan:

docker run --rm -v $(pwd):/zap/wrk/:rw -t zaproxy/zap-stable \
  zap-baseline.py -t https://staging.example.com -r report.html -I

The baseline scan spiders briefly and runs passive rules only, finishing in a few minutes, which makes it suitable for every pull request. Full active scans belong on a nightly or weekly schedule against a dedicated test environment. Alerts can be tuned to ignore, info, warn, or fail individually in a configuration file, which is essential — the default rule set produces findings that no team will fix, and an ignored report protects nothing.

What It Will Not Find

ZAP is very good at finding mechanical, pattern-matched flaws across a wide surface and completely blind to logic. It cannot tell that an order endpoint accepts a negative quantity, that one customer’s identifier returns another customer’s invoice, or that a discount code can be reused indefinitely. Those are business logic flaws, and finding them requires a tester who understands what the application is for.

Treat automated scanning as the floor rather than the assessment. At Privia Security, we combine tuned automation with manual testing in our web application security services.

FAQ – Frequently Asked Questions

Is ZAP as Good as a Commercial Scanner?

For passive checks, spidering, and proxy work it is competitive with anything. Commercial scanners generally handle complex authentication, modern JavaScript applications, and crawl coverage with less configuration effort, and they tend to produce fewer duplicate findings. ZAP closes much of the gap once properly configured, which is a real cost in engineering time rather than licence fees.

Can I Run ZAP Against Production?

Run passive scanning against production safely, because it only observes traffic you were generating anyway. Active scanning sends attack payloads that create records, trigger notifications, and occasionally cause outages, so it belongs in a test environment with representative data. If production really must be scanned, get written authorisation, agree a window, and notify the operations team first.

Why Did My Scan Only Find Problems on the Login Page?

Almost always because ZAP was not authenticated, so it never reached the application behind the login. Configure a context with authentication details, session management, and a verification request so ZAP can detect a logged-out state and recover. Then check the site tree to confirm authenticated pages were visited before you draw any conclusion from a short report.

How Should ZAP Fit Into Continuous Integration?

Run the packaged baseline scan on every pull request against a deployed preview or staging instance. It takes a few minutes, uses passive rules only, and catches regressions in headers, cookies, and configuration. Schedule a full active scan nightly or weekly against a dedicated environment. Tune the alert thresholds first, or the pipeline will fail on findings nobody intends to fix.

Related Terms

You May Be Interested In These

What Is Loyalty Programme Abuse?

Cybersecurity Terms

17 November 2026

What Is Loyalty Programme Abuse?

Read More
What Is an Attack Vector?

Cybersecurity Terms

17 November 2026

What Is an Attack Vector?

Read More