Browse every security term
Full definitions for SSRF, XSS, JWT bypass, OAuth 2.0, mTLS, and the rest of the vocabulary used on this page.
Resources
Pentrova organises the web, API, and infrastructure flaws we detect into five taxonomic classes: Injection, Access Control, Cryptography, Infrastructure, and Application Security. The classes are deliberately broader than CWE so one class can map cleanly to the coverage that hunts it, and narrower than OWASP Top 10 so each entry points at a concrete detection and exploitation technique.
Every entry below describes the class, the signal our agents look for, and the shape of the deterministic proof the platform ships when a finding is confirmed. Specific catalog sizes and coverage inventories are available to evaluators under NDA via the product console and Trust Center.
The database is curated, not exhaustive. A full CWE cross-walk lives inside the product where it can stay accurate as the catalog changes. The entries here are the ones we expect a security engineer to recognise at a glance while evaluating Pentrova on a first read.
Every detection ends at the same artifact: a replayable PoC bundle that an engineer, an auditor, or an incident responder can verify without calling us. That discipline is what makes the taxonomy useful — each class is a bet on producing deterministic evidence, not probabilistic severity.
Browse every security term
Full definitions for SSRF, XSS, JWT bypass, OAuth 2.0, mTLS, and the rest of the vocabulary used on this page.
Unsafe composition of attacker-controlled input into commands, queries, templates, or markup.
Parameter concatenation into SQL statements enables data extraction, authentication bypass, and file reads.
Reflected, stored, or DOM-based injection that reaches a scripting sink in the victim origin.
User input evaluated by a server-side template engine; frequent path to remote code execution.
XML parsers dereferencing external entities, enabling file reads, SSRF, or denial of service.
Missing or inconsistent authorisation checks that let one principal act on another principal’s resources.
Object identifiers accepted from the client without a server-side ownership check.
The API-native sibling of IDOR; top item in the OWASP API Security Top 10.
Weak JWT validation — `alg=none`, HS256/RS256 key confusion, unverified `kid` headers.
Cryptographic primitives used incorrectly: weak algorithms, predictable randomness, or misconfigured TLS.
Server-side acceptance of expired, self-signed, or CN-mismatched client certificates.
Reflected `Access-Control-Allow-Origin` combined with `Allow-Credentials: true`; adjacent to the CSRF class.
Server-side requests, file reads, and remote code execution exposed through infrastructure primitives.
Application fetches URLs supplied by the attacker; pivots into internal services and cloud metadata.
Arbitrary command or code execution on the target — canonical example is the Log4Shell chain.
JNDI-triggering log strings reach a vulnerable Log4j 2.x version and load attacker-controlled classes.
Business-logic and application-layer flaws that are specific to how the service composes its own features.
Correctness depends on the relative timing of concurrent requests — classic time-of-check-to-time-of-use.
A malicious origin induces a browser to send an authenticated state-changing request to a target site.
Misconfigured redirect URIs, missing PKCE, and state-parameter drift in delegated authorisation flows.
Workflows enforced in the UI but not on the server — step skipping, coupon reuse, negative-quantity purchases.
Every class above is backed by dedicated coverage. Book a demo and we will walk the catalog end-to-end against a target of your choice.