Phaerixa Discovery Forensics & Indexing

Changelog

1.3.3

• Security: internal hardening.

1.3.2

• Fixed: a Quick Edit (a new date, status or category) marked the SEO score in the posts list as out of date even though nothing it measures had changed. The score now only goes out of date when the title, content, excerpt, slug or featured image change.
• Changed: Phaerixa’s menu sits below Comments, beside the other Rob Art tools, instead of at the end of the admin menu.

1.3.1

• Changed: Phaerixa’s menu no longer asks for a high place in the admin menu. WordPress now lists it after its own items (Posts to Settings).
• Improved: on the Dashboard, an odd last statistic box spans the full width instead of leaving a gap.
• Fixed (Pro): Search Console’s Re-sync never updated the page list. It now re-checks every page against Google and updates the ones that changed.
• Fixed (Pro): a page was only matched to Google’s results when its address was written exactly the same way. Small differences such as a trailing slash or www no longer stop the match.
• Fixed (Pro): the summary cards grouped pages by severity, so pages Google knows about but has not yet shown in search were counted wrongly. They now have their own “No impressions” group, and clicking a card shows just those pages.
• Fixed (Pro): the Status sort on the Search Console list did nothing. It now sorts by group, with valid pages first.
• Fixed (Pro): when Google returns no data, Phaerixa now says so and names the Search Console property it asked, and the report runs up to today.
• Improved (Pro): a “No impressions” slice in Statistics and a matching figure on the Dashboard.

1.3.0

• Changed: Google Search Console is now part of Phaerixa Pro. The free version contains no Google connection and contacts no outside service — it only reads your own site. Pro installs keep their connection.
• Changed: features that belong to Pro now live entirely in the separate Pro download instead of being switched off inside this one. Where one would appear, the free version shows a short description of what Pro adds.
• Fixed: Statistics → Performance (Pro) showed zeros when Google returned no data or refused the request. It now says which, in words, with Google’s own message. The figures show “—” instead of 0, and a Connection details panel shows exactly what Google answered, with a Refresh from Google button.
• Fixed: on a site whose homepage lists the latest posts, the Homepage description set in SEO → Meta & Social was saved but never added to the page, while the scanner reported it as set. It is now printed, together with the homepage’s social-card title and description.
• Fixed: the scanner now checks the robots.txt and sitemap that crawlers actually receive, including a robots.txt file placed in the site root. It previously looked only at text saved in Phaerixa.
• Fixed: a static front page no longer counts as “missing a meta description” when the Homepage description covers it.
• Fixed: the image alt-text figure could count alt text left behind by deleted images.
• Fixed: the same score could read “Good” on one screen and “Needs attention” on another. Every score now uses the same bands: 80 and above is Good, 50 and above Needs work.
• Fixed: the setup wizard asked for a licence key in the free version and promised “one-click fixes” and “guided repairs”, which Phaerixa does not do. It now describes what Phaerixa does and what Pro adds.
• Fixed: the readme and upgrade panels advertised an “auto-fix engine” that does not exist. Removed.
• Changed: robots.txt snippets added by Phaerixa are now marked with its current name. Snippets marked with its old name are still recognised.

1.2.1

• Security: the stored licence record is signed so it cannot be edited by hand to grant Pro features. That signature used a fixed value shipped inside the plugin, which meant it was the same on every installation and offered no real protection. Each site now generates its own private signing value, which never appears in the plugin’s files. Pro users will need to re-enter their licence key once after updating.

1.2.0

• New: an Upgrade page, reached from its own menu item, setting out what Phaerixa Pro adds and what it costs. The free version previously had no way to see this.
• New: “Upgrade to Pro” buttons throughout the plugin now open a details panel. They previously did nothing at all when clicked.
• New: direct links into your own Google Search Console, from the sidebar on every page, from the Search Console tabs, and — most usefully — from a page’s verdict in the Investigator, which opens Google’s URL Inspection for that exact page.
• Fixed: Phaerixa reported that robots.txt was blocking Google on sites where it was not. Blocking AI crawlers such as GPTBot is a common and harmless practice, and the check mistook it for a site-wide block — then presented that as the decisive reason a page was not indexed. The rule is now read for the specific crawler it applies to.
• Fixed: the Sitemap panel claimed Phaerixa was generating your sitemap and listed addresses that did not exist. Phaerixa does not generate sitemaps; it now identifies the sitemap your site actually publishes, confirms it works before saying so, and reports what it tried when none can be found. This also fixes “Submit to Google” failing.
• Fixed: robots.txt and the Dashboard no longer advertise a sitemap address that redirects, which Google rejects.
• Fixed: a change in the sitemap check meant a working sitemap could be recorded in the Change Ledger as missing.
• Fixed: the Honeypot could write an unchecked address into the server configuration file, which on some hosts could take the whole site offline. Addresses are now validated first and the rule is written in a form older servers also understand.
• Fixed: the Honeypot list no longer runs long browser identifiers across the neighbouring columns; they are shortened, with the full text on hover.
• Fixed: the Change Ledger recorded “Noindex changed” entries where nothing had actually changed, reading “Indexed → Indexed”. A page that was already indexable could be written two different ways internally, which slipped past the check that suppresses non-changes.
• Changed: pages with no Google Search impressions are now labelled “No search impressions” rather than “Not seen by Google”, and treated as information rather than a warning. The old wording implied Google could not see the page, when it only meant the page had not yet appeared in search results — which is normal for anything newly published, and for every page while Google is still processing a site’s data.
• Security and compliance: output escaping tightened across the admin screens, all inline styles and scripts moved into proper files, request data consistently sanitised, and the documentation of external services expanded to name every Google address the plugin contacts. A privacy section has been added covering what the honeypot stores.

1.1.12

• Fixed: the honeypot now records the real visitor IP address behind a reverse proxy or managed host. It previously preferred the server’s own connection address, so on proxied hosting a triggered honeypot could log and block the proxy rather than the offending bot.

1.1.11

• Fixed: the Statistics “Issues Breakdown” chart now shows your current live issue counts, matching the Diagnostics page exactly. It previously read the last recorded scan snapshot, so it could show issues you had already fixed.
• Fixed: the Statistics “Trends” tab now always shows your real recorded scan history instead of illustrative demo data, so figures like “Best Score Recorded” can no longer contradict your live score. When too few real scans have been recorded yet, it shows an honest “trends build up as your site is scanned” message rather than a misleading one- or two-point line. Added axis labels to the trend charts.
• Fixed: the Meta & Social “Meta Coverage” panel now counts coverage the same way the Diagnostics scanner does — it credits per-type default title/description patterns and excludes noindex pages — so it no longer reports pages as “missing” that Diagnostics correctly treats as covered.

1.1.10

• Added: a “This page isn’t an article” toggle in the SEO panel (Advanced). Tick it on a shortcode, tool or landing page whose value is its function rather than long text, and the word-count / thin-content and readability checks are skipped for that page — so a short utility page is no longer flagged as thin content.

1.1.9

• Fixed: the plugin dashboard’s on-page issue counts and health score now match the Diagnostics page exactly. The dashboard was using its own separate calculation that didn’t share the diagnostics scanner’s rules (it counted media attachments, ignored per-type defaults, and used a different score formula), so it reported higher “missing title/description” numbers and a different score. Both views now come from the same scan.

1.1.8

• Fixed: the orphan-pages check now recognises internal links written as relative paths (e.g. “/copyright-policy/”), not just full URLs. WordPress’s editor and many themes store internal links relative to the site root, so a page you had genuinely linked to could still show as having no links pointing to it. It now matches both forms.

1.1.7

• Fixed: the orphan-pages check no longer flags blog posts that are listed on your Posts (blog) page or in a category/tag archive — those posts are reachable through those archives, so they aren’t orphaned. Combined with the previous release, the check now reflects pages that genuinely have no internal links pointing to them.

1.1.6

• Fixed: the “pages with no internal links” (orphan) check no longer reports pages that ARE linked. It previously only searched post content, so pages linked from a menu, an Elementor layout, or a custom post type’s archive listing were wrongly flagged. It now recognises custom post type items shown on their archive as linked, and also checks navigation menus and Elementor content — so the count reflects genuinely orphaned pages.

1.1.5

• Fixed: the dashboard’s 28-day Clicks & Impressions chart now plots real Search Console data. It previously generated a plausible-looking daily curve from the 28-day totals, so a connected site saw invented daily peaks and troughs presented as its own performance. The chart now draws nothing when there is no data rather than filling the gap.
• Fixed: duplicate meta title and description checks compare what each page actually renders, not the stored template. Pages sharing a title pattern such as “%title% %sep% %sitename%” were reported as duplicates even though every rendered title was unique. This also catches genuine duplicates the old comparison missed, where a literal value and a template resolved to the same text.
• Fixed: the attachment-pages warning now reads whether attachment pages are enabled before reporting them as indexable. It previously counted attachments and warned regardless, so the notice could not be cleared — and its fix link pointed at Reading Settings, which has no attachment control. A setting that genuinely turns attachment pages off has been added.
• Fixed: the tag-archive warning can now be resolved. It asserted that archives were publicly indexable without checking, so noindexing every archive left the count unchanged. A tag-archive noindex setting has been added, the check reads it, and the detail text now reports how many tags are used on only one post.
• Fixed: images marked as decorative are no longer reported as missing alt text. An explicit empty alt attribute is the correct way to mark a decorative image, and role=”presentation” and aria-hidden say the same thing; all three are now excluded from the tally instead of counted as failures.
• Fixed: references to deleted images are no longer counted as missing alt text. A widget pointing at an attachment that no longer exists is a broken image, not an alt-text problem, and there is no media entry on which the alt text could be set.
• Fixed: images with no source attribute, inline data URIs and SVG icons are excluded from the alt-text tally. These are template scaffolding with no media-library entry to caption.
• Fixed: the SEO score column in the posts list now shows when its number is out of date. The score is calculated in the editor and saved with the post, so any later change made elsewhere left the column showing an older figure with nothing to indicate it. A stale score is now faded and marked, and says so on hover.

1.1.4

• Fixed: the sitemap check no longer reports “No sitemap found” when your sitemap is actually working. It now follows redirects (a /sitemap.xml that redirects to WordPress’s /wp-sitemap.xml is a valid sitemap), checks the core sitemap URL directly, and retries once so a momentary hiccup on the live check can’t trigger a false “missing sitemap” critical.

1.1.3

• Fixed: pages whose body is a shortcode or page-builder layout (for example a help page built from [plugin_help], or an Elementor page) are no longer wrongly flagged as thin content. The scan now measures the rendered word count, so a page that displays plenty of content is judged on what visitors actually see, not on the few words stored in the editor.
• Changed: when you set a default title or description for a content type (SEO → Meta & Social → per-type defaults), every item of that type now counts as having meta in the site health score — matching what is output on the front end. Setting one default for a whole type (e.g. all plugins) now improves the score for all of them at once, instead of each still showing as missing.

1.1.2

• Fixed: the site health scan no longer counts media attachments (or any content type hidden from search) when checking for missing meta titles and descriptions. Attachment pages never carry editorial meta, so including them flooded the score with issues that could never be fixed and could push it to zero. The scan now focuses only on content types that are actually meant to be found in search.

1.1.1

• Fixed: an important settings bug where saving one card on the SEO → Meta & Social (or Sitemap/Schema) screen could silently wipe settings you had saved on another card — for example your homepage title and description, or your schema social links, disappearing after saving a different section. Each setting now keeps its value when you save a different card. If you were affected, simply re-enter and save the cleared fields once; they will now stick.

1.1.0

• Added: SEO for custom post type archives. Archive pages — for example your Books, Portfolio or Products listing — now get their own meta title, description and Collection Page structured data, generated for any custom type automatically. Previously only individual posts and pages were given meta tags.
• Added: per-content-type SEO defaults. In SEO → Meta & Social you can now set a default title and description pattern for each custom content type (e.g. every Book), so a whole type shares a format without editing each item. A title or description set on an individual item still takes priority.
• Added: a Demo data switch (Settings → General, administrators only) that fills the dashboards and trends with sample Search Console data for previews, then turns off to show your real data. Replaces the previous developer-only constant and keeps every screen consistent.
• Changed: the site health scan now honours your “Post types to analyse” choice, so custom types such as Portfolio or Products are included in the score — not just posts and pages.
• Fixed: the page title is now produced through WordPress’s own title system, so there is never a duplicate title tag on themes that manage their own titles.
• Fixed: pages set to noindex are no longer counted as missing a meta description or title, and structured records are no longer flagged as thin content — both were unfairly lowering the score.

1.0.24

• Fixed: the content-light detection from 1.0.22 was too aggressive — pages that DO have a real description stored in a plugin field (portfolio items, book pages, custom listings) were wrongly treated as having no content, so their content and readability checks were skipped. Phaerixa now decides this from the actual amount of readable text on the rendered page, so only genuinely structured records (e.g. a copyright-evidence entry with no public description) skip the article checks; described pages are analysed in full again.

1.0.22

• Added: pages that carry their information in structured fields rather than body text — for example a copyright-evidence record, a directory entry or a gallery item — are now recognised automatically. The content and readability checks that only make sense for an article (word count, headings, in-body keyword, internal/outbound links) are marked “not applicable” and no longer drag the score down, so a well-optimised structured page can reach a high score on the checks that actually apply (title, description, canonical, indexability, schema, images).
• Fixed: the admin styles now load on every Phaerixa sub-page, not just the dashboard.

1.0.21

• Fixed: the per-post SEO checks now work on content types whose body text lives in plugin or theme fields rather than the main editor — for example portfolio items, book pages and custom listings. Phaerixa reads the actual rendered page, so the focus-keyword-in-intro, outbound and internal link, heading, and readability checks register instead of sitting blank.
• Note: after updating, re-save an affected page once so its content is re-analysed.

1.0.20

• Changed: the SEO panel now sits at the bottom of the editor’s metabox column instead of the top, so your own page panels appear first. (You can still drag it wherever you prefer; WordPress remembers your order.)
• Housekeeping: aligned the plugin slug and text domain with the WordPress.org naming requirements. No change to your data or settings.

1.0.19

• Fixed: the “Images have alt text” check no longer reports missing alt on Elementor pages where every image actually has alt. Page-builder images are often rendered without WordPress’s wp-image-ID class, so they were being counted a second time as anonymous “no alt” images; they are now matched back to their media item by their image URL (and decorative CSS background images are correctly excluded).

1.0.18

• Fixed: the content “Images have alt text” check now recognises the featured image and images placed with Elementor (and other page builders) — previously these were invisible to the analyser, so a page with every alt set could still read as a warning. Their alt text is now resolved from the media library and folded into the check.
• Fixed: content analysis inside Elementor now sees links, headings and images — the live extractor returned plain text (dropping every link, heading and image), so the internal/outbound link, heading and image checks failed on Elementor pages that actually had them. It now reads the rendered markup, and also handles Elementor’s flexbox‑container layouts where the older widget selector matched nothing.
• Fixed: the SEO meta “coverage” percentages no longer exceed 100% — the total and the “has custom meta” counts now measure the same set of post types (previously products and other types inflated the numerators against a post/page‑only total, e.g. 343%).

1.0.17

• Fixed: a blank title separator no longer breaks title output — the %sep% token, the SERP preview and the active separator option now all fall back to the default when the stored value is empty
• Fixed: the SERP title length bar now fills correctly at the ideal length (full at the optimal maximum, pinned and recoloured past it) instead of reading half-full
• Added: Pro updates are delivered directly from the developer’s licence server — a licensed Pro install now receives update notices in place (the update link lights up and installs the new version)

1.0.16

• New Decision Path in the Investigator — see the exact reasoning behind a page’s verdict as a step-by-step flow (reachable, published, crawlable, index allowed, canonical, Google’s verdict), with the deciding step highlighted and a one-click link to where you fix it
• New Index Matrix — a site-wide table of every published URL with at-a-glance signal columns and an indexability verdict, filterable by type and verdict, each row opening its full dossier
• Dashboard now shows an Indexability summary (indexable / canonicalised / blocked across your whole site) and, for Pro, a Fix Verification tally (awaiting / recovered / regressed)
• Fixed: turning a page noindex now correctly logs as an indexing concern and appears on the dashboard and Change Ledger (previously some meta changes were silently missed)
• Fixed: a fix is only reported as “Recovered” once genuinely confirmed — an indexing fix without Search Console connected now honestly reads “Awaiting Google” rather than over-claiming
• Various Investigator layout and wording polish

1.0.15

• New Investigator workspace — search any URL on your site and open its full “dossier”: a plain-English indexing verdict plus an Evidence Stack showing every signal (HTTP status, robots.txt, meta robots, canonical, sitemap, internal links, publish status and Google’s own verdict) that decides whether the page gets indexed
• Type-ahead search — start typing and Phaerixa suggests your own pages, posts and products by name, so you pick a URL instead of copy-pasting it, plus a Clear button to start a fresh search
• Verify results (Pro) now show right in the Change Ledger — each fixed change is badged Recovered, Awaiting Google, or Regressed, so you can see at a glance whether a fix actually took
• Layout fix on the Investigator screen

1.0.14

• Google Search Console connection is more reliable — after connecting, Phaerixa now automatically selects and saves the right property for your site, and the status now reflects a genuinely working connection rather than just a stored token
• Verify (Pro) — when you fix an indexing problem, Phaerixa now watches whether Google actually recovers the page, marking it Awaiting Google, Recovered, or Regressed, using Google’s own reporting so “Recovered” means Google really re-indexed it
• Internal reliability and code-quality improvements

1.0.13

• Change Ledger now watches far more — beyond editor changes it also spots robots.txt edits, pages leaving your sitemap, a site-wide noindex header, and per-page HTTP problems (404, server errors, soft-404, forbidden, redirect errors) — the technical reasons Google drops pages, caught the moment they happen
• New “Why this page may not be indexed” panel in the post SEO box — when a page has an active indexing problem it tells you what’s wrong, when it changed, and how to fix it (only appears when there’s something to report)
• Redesigned Dashboard leads with your indexing activity: recent search-facing changes and any active concerns are front and centre, with your SEO health score alongside

1.0.12

• Product SEO scoring — WooCommerce products now use a product-appropriate profile: word count and keyword checks combine the long and short descriptions, image alt-text scanning includes the product featured image and gallery, and article-only checks (headings, internal/outbound links) no longer count against product pages
• Product schema now emits the fields Google recognises for a rich result — Offer (price, currency, availability), SKU, short description, and star rating — pulled from WooCommerce
• SEO title and meta description character counters and length checks now measure the resolved value (e.g. %title%) rather than the literal template text
• Info tooltips in the post SEO panel now appear at the button you click, not a fixed corner
• Posts-list SEO score badge now matches the score shown in the post editor
• Change Ledger (new) — records search-relevant changes (indexing, canonical, slug, status, permalinks, search visibility, plugin and theme changes) so you can see what changed and when
• Housekeeping — bot honeypot is now fully Pro-only with no trace in the free build; added a distribution manifest so only shipping files are packaged

1.0.11

• WordPress 7.0 compatibility — restored familiar form, button, checkbox and radio styling that WP 7.0 changed
• Added empty-state overlay on Honeypot trap activity chart for sites with no data yet
• Fixed honeypot chart axis labels not rendering (CSS variable typo)

1.0.0

• Initial release — diagnostics engine, full core SEO layer, GSC integration, per-post panel, honeypot bot trap, setup wizard