Fifteen Coder fingerprints and 7,392 title matches describe very different evidence
A scoped explanation of why Coder application and title counts differ by orders of magnitude and what to verify on owned systems.
Tag archive
A scoped explanation of why Coder application and title counts differ by orders of magnitude and what to verify on owned systems.
A method note on the large difference between ntfy title and application fingerprints and the access checks neither count replaces.
A small population makes individual review practical. Three questions for a self-hosted chat server, and how version information turns measurement into a maintenance check.
When the cluster console is published through an ingress, the API server becomes internet-facing. Four checks that decide whether that is acceptable.
A worked example showing how query selection changes an exposure count by orders of magnitude, using the SonicWall SMA1000 vulnerabilities.
A large title count and a four-match fingerprint describe a platform whose administration page is recognisable but whose device signature is not. What that means for local checks.
A repeatable method for building an internet-exposure baseline with ZoomEye so that vulnerability scoping takes minutes instead of hours.
Why a zero result from ZoomEye's CVE-indexed field is a coverage signal rather than a population count, and how to test the field properly.
Measuring the SonicWall Edge: 15,608 WorkPlace Interfaces and the Limits of That...
A ZoomEye query returned 9,941 reachable N-central interfaces. Here is what that number does and does not tell defenders about the CVE-2026-86218 attack surface.
IoT botnet notices keep describing Telnet brute force because the reachable population never closed. Internet measurement shows the entry path and its limits.
Counting the AI Gateway Surface: ZoomEye Data on LiteLLM and Kestra After the September KEV...