Nikto is an open-source web server scanner. It sends a large number of requests to a target web server and checks the responses against a maintained database of tests: default and sample files left in place, dangerous or unnecessary HTTP methods, outdated server and interpreter versions, directory indexing, backup and configuration files exposed under predictable paths, and missing or misconfigured security headers. It is a fingerprinting and misconfiguration tool rather than a deep application scanner, and that distinction defines both what it is good at and what it should never be relied on for.
What Nikto does well is breadth at speed. Pointed at a host, it will enumerate the server software, probe several thousand known-bad paths and report the ones that respond, usually within minutes. That makes it valuable as a first-pass sweep across an estate of web servers, as a regression check that a hardened build has not drifted, and as the fast layer in front of slower, deeper dynamic testing. What it does not do is understand your application: it will not authenticate through a modern login flow, follow a JavaScript-rendered route, or reason about business logic.
Nikto is deliberately noisy. It does not attempt to hide, it will generate a large volume of 404 responses in the target's logs, and it will trip a well-configured WAF or intrusion detection system almost immediately. That behaviour is genuinely useful in a defensive context — running Nikto against your own estate is a reliable way to confirm your detection and blocking actually work — but it also means scanning is something you do against systems you own, with authorisation recorded, and with a plan for what the target's logs and alerting will do while you run it.
Why this skill matters now
Estates accumulate web servers that nobody owns. Reverse proxies, admin consoles, monitoring interfaces, legacy applications behind a load balancer, forgotten staging hosts — each one is a web server with a default configuration until somebody checks. Cloud and container adoption made spinning one up trivial and made keeping track of them harder, so periodic sweeping became a real operational need rather than a formality.
Nikto stays relevant because it is fast, free, scriptable and covers exactly the category of problem that keeps appearing in incident reports: an exposed configuration file, an admin interface reachable from the internet, a server banner revealing an unpatched version, a directory listing that should have been off. Those are cheap to find and cheap to fix, and they are found by breadth rather than depth.
The skill is in using it responsibly and usefully. Anyone can run the default scan; the difference is knowing how to tune it so a sweep across two hundred hosts finishes overnight, how to chain it through a proxy so results feed a wider workflow, how to separate the genuinely actionable findings from a long list of informational noise, how to verify a finding before escalating it, and how to run it as a scheduled control against your own infrastructure without paging the security operations team every week.