Audit response + SPRINT-001: disprove SEO claims, then fix the copy they were right about

.okf/build/hugo-build.md

@@ -7,6 +7,7 @@ tags: [build, hugo, validation]
 generated:
   by: process:okf-migrate
   at: 2026-07-19T12:00:00Z
+timestamp: 2026-08-13T00:00:00Z
 ---
 
 `bin/hugo-build` builds the site into `_dest/public-dev/` (repo-root
@@ -41,3 +42,34 @@ safelist entries). Guards: bin/hugo-build runs a warm-up pass when
 production + stats missing; the deploy workflow has an explicit warm-up
 step. sr-only/skip-link also safelisted as defense-in-depth. Never
 trust a first cold production build's CSS.
+
+# Minified output has unquoted attributes (2026-08-13)
+
+`minifyOutput = true` makes Hugo drop quotes on attribute values with no
+spaces: `rel=canonical`, `name=description`, `type=application/ld+json`.
+Valid HTML5, and Google parses it correctly - but regex-based third-party
+SEO/AEO audit tools require quotes and report the site as missing canonical
+tags, meta descriptions, and structured data. All three are false.
+
+Diagnostic tell: checks reading ATTRIBUTE VALUES fail while checks reading
+ELEMENT CONTENT (title, H1) pass. That split means parser artifact, not site
+defect. Settle it in one call with GSC `inspect_url_enhanced`, which returns
+`user_canonical` and the rich-results verdict - Google reporting what it
+actually parsed.
+
+Note `config/test/hugo.toml` sets `minifyOutput = false`, so **the test suite
+never sees minified output** and cannot catch minification-related regressions;
+that needs a `hugo --environment production` build. Full write-up:
+`docs/projects/2510-seo-content-strategy/seo-review-2026-08-13.md` §8.
+
+**Our own suite had the same defect** (2026-08-14). Turning `keepQuotes` on
+broke three tests in `test/integration/hugo_pipeline_test.rb` (the CI "Asset
+Pipeline" job) that matched unquoted attribute literals -
+`include?("crossorigin=anonymous")`, `rel=stylesheet`, `as=style`. The
+`integrity="sha256-..."` assertions kept passing because base64 forces quotes
+either way. Same tell as the audit tool: attribute-VALUE matches break,
+element-content matches survive. Fixed with an `attr(name, value)` helper
+matching either form - assert the shape, not the minifier setting. If you
+touch minify config, run `bundle exec ruby -Itest
+test/integration/hugo_pipeline_test.rb` (~12s local, ~568s on CI because it
+runs two full Hugo builds).

.okf/build/index.md

@@ -1,6 +1,6 @@
 # Build & Test
 
-* [Hugo build pipeline](hugo-build.md) - bin/hugo-build with the 8 course validators
+* [Hugo build pipeline](hugo-build.md) - bin/hugo-build with the 8 course validators; also the PurgeCSS cold-start race and the minified-unquoted-attribute audit-tool trap
 * [Test gates](test-gates.md) - the local suites and when each is a commit blocker
 * [CI gates](ci-gates.md) - what GitHub Actions enforces: build, unit, path-scoped link check (visual regression is report-only), and what gates a PR never sees
 * [Template PDFs](pdf-templates.md) - regenerating the downloadable course PDFs

.okf/build/test-gates.md

@@ -28,6 +28,23 @@ extend it when adding components or critical files. The macOS full suite remains
 
 # Hard-won caveats
 
+- **The 2% default tolerance hides small text/colour changes** (2026-08-14).
+  `DEFAULT_SCREENSHOT_CONFIG = {tolerance: 0.02}`
+  (`test/application_system_test_case.rb:87`). Turning a four-word phrase into
+  a link on the homepage changed ~0.24% of the frame, so the gate PASSED and
+  the baseline was never re-recorded - it still shows the pre-change render.
+  Consequence: a green visual suite does NOT mean "no visual change", only "no
+  change larger than 2% of the frame". For link/colour/short-text edits,
+  verify by reading the built HTML or the render, not by trusting green. This
+  is the false-green class documented in
+  `docs/20-29-testing-qa/test-architecture-anti-masking.md`.
+- **`FORCE_SCREENSHOT_UPDATE=1` re-records EVERYTHING** (2026-08-14). On
+  `bin/dtest` it also disables the `git checkout -- .../linux` guard that
+  normally discards sub-tolerance Rosetta drift, so a run rewrites all 45
+  Linux baselines rather than the few your change moved. Procedure: run it,
+  copy out only the baselines your change legitimately moved, `git checkout --
+  test/fixtures/screenshots/linux`, then copy your files back. On `bin/qtest`
+  the flag appears to be ignored entirely - the suite still compares.
 - **A `skip_area` selector that matches NOTHING costs 5s per screenshot**
   (2026-08-01). snap_diff resolves each mask via `all(sel, visible: true)`,
   and Capybara waits `default_max_wait_time` (5s) on a zero-match selector.

.okf/log.md

@@ -1,5 +1,27 @@
 # Bundle Update Log
 
+## 2026-08-14 (claims) - the number nobody had a source for was wrong
+
+* **Update**: `.okf/build/test-gates.md` gains the 2% tolerance false-green and
+  the `FORCE_SCREENSHOT_UPDATE` re-record trap;
+  `docs/projects/2510-seo-content-strategy/seo-review-2026-08-13.md` §8 actions
+  7-9 closed.
+* **The finding**: the site published "4.8/5 by 32 clients" and
+  `reviewCount: 32` in schema on ~1,147 URLs. `reviewCount` had **no source
+  anywhere in the repo** - `data/company.yaml:11` cites a Clutch rating with no
+  count. The live Clutch profile shows **4.8 from 9 reviews**. The rating was
+  right; the count was overstated ~3.5x. `data/course_banned_strings.yaml:65`
+  had already banned "4.8/5" in course content as a "volatile third-party
+  review score" - the course side learned this and the marketing side did not.
+* **The rule**: a number with no in-repo provenance is a defect, not a detail.
+  When a claims audit says "verify X", verify it before deciding what to do
+  with it - the earlier call to keep 32 was made assuming it was sourced.
+  Prefer a **linked** rating over a bigger unlinked one; the link is the proof.
+* **Also closed**: fabricated `Review` objects ("Technology Executive",
+  "Startup Founder") deleted from `comprehensive-service-schema.html`, and
+  `keepQuotes = true` added (verified in a production build; homepage
+  124,256 -> 125,988 bytes, +1.4%).
+
 ## 2026-08-13 (LinkedIn exhibits) - purpose-built post images consume the house spec
 
 * **Update**: `linkedin-posts/README.md` §"Every post carries a visual" rewritten
@@ -24,6 +46,29 @@
   MBP-14), image column 460px, image is a click-to-open-new-tab link
   (drag-to-attach preserved); prev/next nav now traverses in board order
   (chronological by effective date), not Hugo section order.
+## 2026-08-13 (build) - a failing audit tool is not a failing site
+
+* **Update**: `.okf/build/hugo-build.md` gains "Minified output has unquoted
+  attributes"; full write-up in
+  `docs/projects/2510-seo-content-strategy/seo-review-2026-08-13.md` §8.
+* **The finding**: a third-party AI-SEO tool scored the site 60/100 "Multiple
+  Organ Failure" and reported no canonical tags, no structured data, and
+  placeholder meta descriptions. **All false.** `minifyOutput = true` drops
+  quotes on space-free attribute values (`rel=canonical`,
+  `type=application/ld+json`), and the tool's regex parser requires quotes.
+  GSC confirms Google parses all of it correctly.
+* **The tell**: checks reading ATTRIBUTE VALUES failed; checks reading ELEMENT
+  CONTENT (title, H1) passed. Read the failure *shape* before the failure
+  *text* - that split is a parser artifact every time.
+* **The rule**: before acting on any external audit, run GSC
+  `inspect_url_enhanced`. It returns `user_canonical` plus the rich-results
+  verdict - Google reporting what it actually parsed - and settles it in one
+  call. Also note `config/test/hugo.toml` sets `minifyOutput = false`, so the
+  test suite is blind to minification regressions by construction.
+* **What was real**: the same audit's copy criticism. And what it missed
+  entirely - the site publishes fabricated review schema (invented "Technology
+  Executive" reviews + two contradictory ratings). Asking "is schema missing?"
+  never asks "is the schema honest?".
 
 ## 2026-08-13 (visual gate) - a new component needs cold eyes, not the implementer's
 

config/_default/hugo.toml

@@ -63,6 +63,13 @@ disableKinds = []
 
 [minify]
   minifyOutput = true
+  # Regex-based third-party audit crawlers cannot match unquoted attributes
+  # (rel=canonical, name=description, type=application/ld+json), and report the
+  # site as missing canonical tags and structured data. Both are present and
+  # Google parses them fine - see seo-review-2026-08-13.md §8. This is a
+  # cosmetic hedge for those tools, NOT an SEO fix.
+  [minify.tdewolff.html]
+    keepQuotes = true
 
 [[redirects]]
   from = "/tags/:slug/"
@@ -111,47 +118,47 @@ disableKinds = []
       parent = "services"
       identifier = "fractional-cto"
       title = "Fractional CTO"
-      name = "Get on-demand access to a CTO to help guide your technical vision, accelerate team-building, and improve development team operations."
+      name = "A senior engineer reviews your codebase and your team, then tells you in plain English what to fix first. $5K-$15K/month, starting in days."
       pageRef= "/services/fractional-cto/"
       weight = 1
 
     [[menu.main]]
       parent = "services"
       identifier = "fractional-product"
       title = "Fractional Product Management"
-      name = "Get on-demand access to a product expert to help design UX, plan & prioritize your roadmap, and manage development schedules."
+      name = "A product lead who runs your roadmap and writes the specs, so your developers stop guessing what to build next."
       pageRef= "/services/fractional-product-management/"
       weight = 2
 
     [[menu.main]]
       parent = "services"
       identifier = "app-web"
       title = "App/Web Development"
-      name = "Whether you’re a startup building new products or an established business upgrading existing systems, we help deliver positive outcomes."
+      name = "Rails and React builds, test-driven from the first commit. You own the code after every milestone."
       pageRef= "/services/app-web-development/"
       weight = 3
 
     [[menu.main]]
       parent = "services"
       identifier = "outsourced-developer"
       title = "Outsourced Developer Staffing"
-      name = "Increase your development capacity & reduce administrative workloads with pre-trained engineers that are ready to deploy within weeks."
+      name = "Senior Rails and React engineers join your team in weeks. They follow your process, and you keep control of the code."
       pageRef= "/services/outsourced-developer-staffing/"
       weight = 4
 
     [[menu.main]]
       parent = "services"
       identifier = "software-qa"
       title = "Software AQA & Testing"
-      name = "Accelerate development, streamline release cycles, and eliminate roadblocks with fully-managed software testing & QA services."
+      name = "We write the tests your app never had, then send you a bug report you can actually read - what broke and what it costs to fix."
       pageRef= "/services/software-qa-cat/"
       weight = 5
 
     [[menu.main]]
       parent = "services"
       identifier = "talent-recruiting"
       title = "Talent Recruiting & Training"
-      name = "Scale your development team & simplify the talent acquisition process with top-caliber candidates that are pre-vetted by our team."
+      name = "We screen Rails and React candidates the way we screen our own hires - take-home code and a live pairing session. You interview the short list."
       pageRef= "/services/talent-recruiting-training/"
       weight = 6
 
@@ -165,47 +172,47 @@ disableKinds = []
       parent = "usecases"
       identifier = "startup-mvp"
       title = "Startup MVP Prototyping & Development"
-      name = "Whether creating an interactive prototype to validate the user experience or building out a Minimum Viable Product (MVP) for initial customers, we help startups accelerate the path from idea to revenue."
+      name = "A clickable prototype you can test with real users, or a working MVP your first customers can pay for. You see a demo every week."
       pageRef= "/use-cases/startup-mvp-prototyping-development/"
       weight = 1
 
     [[menu.main]]
       parent = "usecases"
       identifier = "salvage-project"
       title = "Salvage an Existing Project"
-      name = "The sooner a software project is salvaged, the better chance it has of being a viable long-term investment. We can help transform your product, providing the visibility, clarity, and control you need to get your project on track."
+      name = "A senior engineer audits your codebase in 48 hours and writes you a recovery plan with costs attached."
       pageRef= "/use-cases/salvage-existing-project/"
       weight = 2
 
     [[menu.main]]
       parent = "usecases"
       identifier = "empower-team"
-      title = "Empower an Existing Engineering Team"
-      name = "The easiest & fastest way to grow software development capacity is by using a trusted technology partner that can deploy pre-trained developers on-demand without the need for additional supervision."
+      title = "Extend an Existing Engineering Team"
+      name = "Add senior Rails and React engineers without adding management overhead. They ship inside your existing process."
       pageRef= "/use-cases/empower-existing-engineering-team/"
       weight = 3
 
     [[menu.main]]
       parent = "usecases"
       identifier = "emergency-cto"
       title = "Emergency CTO Leadership"
-      name = "To maintain visibility, accountability, and control over technology outcomes when a critical team member leaves, companies need on-demand access to technical leadership to help keep the status quo for IT intact."
+      name = "Your lead developer just quit. We step in, take stock of what actually exists, and keep the product shipping."
       pageRef= "/use-cases/emergency-cto-leadership/"
       weight = 4
 
     [[menu.main]]
       parent = "usecases"
       identifier = "automate-product"
       title = "Automate Product QA & Testing"
-      name = "To enable faster product development while eliminating the risks of costly mistakes, companies need access to a reliable product testing & QA resource that can ensure quality at any scale."
+      name = "Automated tests that catch the break before your users do, plus a weekly report on what they caught."
       pageRef= "/use-cases/automate-product-qa-testing/"
       weight = 5
 
     [[menu.main]]
       parent = "usecases"
       identifier = "accelerate-dev"
       title = "Accelerate Development & Maximize Capacity"
-      name = "Our developers are ready to deploy within weeks (not months), giving companies the ability to accelerate their software development process while still maintaining control over engineering quality."
+      name = "Senior engineers start in weeks, not months. Every milestone ships with tests and a written progress report."
       pageRef= "/use-cases/accelerate-development-maximize-capacity/"
       weight = 6
 

content/pages/about-us/index.md

@@ -24,15 +24,15 @@ founder_expertise:
       value: "Our leadership team averages 12+ years of industry experience, with specializations in Ruby on Rails, React, startup MVP development, and fractional CTO services. We've contributed to 50+ open-source projects and published 540+ technical articles sharing our expertise with the developer community."
 
     - name: Industry Recognition
-      value: "Clutch Top Ruby on Rails Developers (2023-2024), featured in Forbes and Inc. Magazine for technical leadership excellence, recognized for 95% client retention rate—highest in the industry for development agencies."
+      value: "Clutch Top Ruby on Rails Developers (2023-2024), featured in Forbes and Inc. Magazine for technical leadership, with 95% of clients continuing past their first engagement."
 
     - name: Proven Track Record
       value: "Delivered 200+ successful projects for startups and growing companies across healthcare, education, SaaS, and e-commerce sectors. Our clients achieve 89% fundraising success rate and 3x faster time-to-market compared to traditional development approaches."
 
 about_us_block1:
   headline: Our Mission
   items:
-    - In the software development world, two world-class developers can outperform an army of sub-par engineers.
+    - Two strong developers will outbuild a room full of average ones, and cost less.
     - At JetThoughts, we believe in the power of curating technical talent and strive to differentiate ourselves on the quality of our work.
     - To ensure we deliver the best outcomes every time, we choose to work with only a few clients at a time and make them our primary focus.
     - By giving each client more attention in the spotlight, we're able to improve alignment, make more progress, and create better long-term relationships.

content/services/app-web-development/index.md

@@ -11,7 +11,7 @@ cover_image: service-app-web-development.jpg
 menu_custom:
   icon: submenu-web.svg
   title: App/Web Development
-  description: Whether you’re a startup building new products or an established business upgrading existing systems, we help deliver positive outcomes.
+  description: Rails and React builds, test-driven from the first commit. You own the code after every milestone.
 
 metatags:
   image: og-services-app-web-development.jpg
@@ -26,7 +26,7 @@ faqs:
   - question: "What's your development process?"
     answer: "We follow an agile development methodology with regular sprint cycles, continuous integration, and frequent client communication. Our process includes discovery and planning, design and prototyping, iterative development with weekly demos, testing and QA, deployment, and ongoing support."
   - question: "Can you work with our existing team or systems?"
-    answer: "Absolutely. We seamlessly integrate with existing development teams, work with legacy systems, and can augment your in-house capabilities. Our team acts as an extension of your organization, adapting to your processes, tools, and communication preferences. If you need to scale your team quickly, our [outsourced developer staffing](/services/outsourced-developer-staffing/) service can provide additional Ruby on Rails developers. We also ensure quality through our comprehensive [software QA testing services](/services/software-qa-cat/) integrated into the development process."
+    answer: "Absolutely. We join existing development teams and work with legacy systems, adapting to your processes, tools, and communication preferences. If you need to scale your team quickly, our [outsourced developer staffing](/services/outsourced-developer-staffing/) service can provide additional Ruby on Rails developers. We also ensure quality through our comprehensive [software QA testing services](/services/software-qa-cat/) integrated into the development process."
 
 overview:
   headline: From idea to production MVP in 8 weeks

content/services/fractional-cto-cost/index.md

@@ -224,6 +224,6 @@ Our fractional CTO cost-effective solutions support businesses across multiple i
 
 ## Ready to Get Expert CTO Leadership at Fraction of the Cost?
 
-Understanding fractional CTO cost is the first step toward accessing world-class technology leadership for your business. Our transparent pricing and flexible engagement models make strategic CTO expertise accessible to companies of all sizes.
+Understanding fractional CTO cost is the first step toward getting senior technology leadership for your business. Our transparent pricing and flexible engagement models make strategic CTO expertise accessible to companies of all sizes.
 
 [Contact us today](/contact/) for a free consultation and customized fractional CTO cost proposal tailored to your specific needs and growth objectives.