Notes & analysis

Blog

Shorter notes between the reports — what the pipeline caught this week, how a detector works, and where the supply chain is heading.

6 posts

04 Nov 2026

The Brotli RCE, by package count: where a vulnerable decoder actually ships

Imperva's Node2Shell (CVE-2026-111497) hits the Brotli C decoder 1.0.3–1.2.0. We mapped which published packages actually bundle a vulnerable copy — the real ones live in RubyGems and crates.io, npm's Brotli packages are pure JS/WASM ports, and at least 8 of the 13 Node runtime versions in our corpus predate the fix.

vulnerabilitybundled dependenciesexposure analysis
8 min · Read →
16 Oct 2026

Joomla's Two Upload Gates: Why PoCs for CVE-2026-48908 and CVE-2026-48939 Fail on Core Checks Nobody Mentions

Two KEV upload flaws in Joomla extensions (SP Page Builder, iCagenda) are neutralized by two undocumented core gates: a case-sensitive archive content scan (shell.PHP gets through) and a case-insensitive extension check (.htaccess gets through). Verified live, with a reproducible harness.

joomlacve-2026-48908cve-2026-48939file-uploadcisa-kevdast
12 min · Read →
15 Oct 2026

When SAST Can't See the Patch

A WordPress-core traversal where the fix is a backported guard: source-diff SAST sees the same code in vulnerable and patched trees, a one-request dynamic oracle tells them apart, and the precondition turns out to be the rule, not the exception.

wordpresslfipath-traversaldastsastcve-2026-87902
7 min · Read →
05 Oct 2026

The libheif RCE, by package: who actually ships a vulnerable copy

Wordfence found a critical RCE in libheif (1.18.0–1.23.2). We mapped which published packages bundle a vulnerable copy — sharp's libvips build shipped one through 1.3.3, some forks and binaries still do, and the package version routinely lies about what's inside.

vulnerabilitybundled dependenciesexposure analysis
7 min · Read →
28 Jul 2026

Why we attribute every indicator to a package

Most threat feeds hand you a list of domains and no provenance. Ours name the package, version and file each indicator came from — so you can check whether it even concerns you, and verify it before you block.

feedsmethodology
5 min · Read →
30 Jun 2026

A false positive is an asset, not an embarrassment

Every confirmed false positive becomes a permanent negative test case. Here is how that feedback loop actually changes detector behaviour — and why we treat clearing a package as a result, not a wasted scan.

methodology
5 min · Read →
Stay current

New editions, in your inbox.

Get notified when codelake Research publishes a new report, threat brief or quarterly advisory roundup. No marketing — just the research.

We use your address only to send research updates. Unsubscribe anytime.