Skip to content

1.3.0: WordPress-aware scanning and honest reporting - #1

Merged
JaydenYoonZK merged 4 commits into
mainfrom
release/v1.3.0
Sep 13, 2026
Merged

JaydenYoonZK merged 4 commits into
mainfrom
release/v1.3.0

Conversation

@JaydenYoonZK

Copy link
Copy Markdown
Member

What this changes

For a WordPress owner

  • A scan recognises the WordPress install, prints its version, and tags flagged files that are the site's own code (config, core, theme, plugin) with advice to replace them with a clean copy rather than only quarantine them. Scanning only wp-content still knows its site.
  • README: a Requirements section, a step-by-step guide to scanning a site on any hosting (over SSH, or on a downloaded copy), and an after-the-scan checklist for the things a file scanner cannot do.
  • jayshield https://example.com now explains that JayShield reads files on disk and shows how to get them, instead of "no such file or folder".

Honest reporting

  • Folders named cache are scanned. wp-content/cache is a common place to plant a shell, and the upload heuristic already treated it as one.
  • The report counts and names the folders skipped by name. --include <dir,...> scans them, --include all scans everything.
  • The clean-scan line now says what a clean result means.

Precision

A pristine WordPress 7.1 download was reported as 26 infected files, 5 critical. All were false positives. The WSO, hex-escape, and hidden-iframe rules and the long-line, high-entropy, and base64-blob heuristics are tightened to what malware actually writes (details in the changelog), and tag attribute scans are bounded so a page of unterminated tags cannot stall a scan.

Clean against WordPress 7.1 (with and without --include vendor), WooCommerce 11.1.0, Astra 4.13.11, jQuery 3.7.1, and Laravel 12.x. Every planted threat in the fixtures is still found. 73 tests.

After merging

  • npm publish from a fresh git pull on main (the package is at 1.3.0).
  • Tag v1.3.0 and create the release from the changelog entry.

…ing clean core

A scan now finds the WordPress install its targets belong to, prints the
version, and tags every flagged file that is the site's own code as config,
core, theme, or plugin, with advice to replace it with a clean copy rather
than only quarantine it. Scanning only wp-content still knows its site.

Folders named cache are scanned. The report counts the folders skipped by
name and lists them, and --include puts them back, with "all" for
everything. A website address is refused with the two ways to scan that
site, over SSH or on a downloaded copy, instead of "no such file or folder".

Rules match what malware writes, not what it resembles: the WSO rule wants
WSO's own function and constant names; the hex rule wants a dangerous
function or request variable spelled in hex, or a hex literal called as a
function; the hidden iframe rule wants a hidden frame with a real
destination and skips the Tag Manager noscript snippet. The long-line,
high-entropy, and base64-blob heuristics also require a decoder, an
execution call, or request input. A cached page whose only PHP is the
die() guard is left alone. Tag attribute scans are bounded so a page of
unterminated tags cannot stall a scan.
…n promise

README gains a Requirements section, a step-by-step guide to scanning a
WordPress site on any hosting with an after-the-scan checklist, the skip
list and --include, and a plain account of what the scanner can and cannot
do, with the clean-code corpus it is checked against. The landing page says
the same in one line.
@JaydenYoonZK
JaydenYoonZK merged commit 1428d56 into main Sep 13, 2026
3 checks passed
@JaydenYoonZK
JaydenYoonZK deleted the release/v1.3.0 branch September 13, 2026 17:54
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant