Clear the rest of the fabricated case studies: ratchet 16 to 9 - #591

Finishes the REPAIR tier. Every actionable carrier the ratchet listed is now
cleared; the 9 survivors are all in the three fractional-CTO posts, parked by an
explicit decision rather than by neglect.

Removed, all invented client work in claims-canon's sense - an anonymous company
carrying precise numbers, with no subject a reader could verify:

- `rails-performance-at-scale` (814 impr): "a real example from our work with a
  fintech startup that grew from 15K to 800K users in 8 months", with
  month-by-month figures. Its "key lessons" went too - five generic scaling
  points the post's own stage sections already make.
- `complete-guide-ruby-rails-ai-integration-2025` (81 impr): 349 lines across
  TWO fabricated sections. Three "Case Study" blocks naming anonymous clients
  with exact revenue ("HR tech SaaS, 15K customers", "$40M ARR"), and the post's
  OPENING section - "Lessons from $180K in Mistakes", asserting outright "these
  aren't hypothetical - they're real projects", with a "$12,000 Weekend" and
  "47,000 GPT-4 API calls". The technical spine survives intact: landscape,
  decision framework, integration patterns, production practices, testing.
- `tdd-workflow-automation-rails-teams`: a case study ending in a fabricated
  **VP Engineering pull-quote**. Invented testimonials are banned by name in
  claims-canon, which makes this the worst item in the batch even though the
  page has no measurable traffic.
- `internal-product-teams-cost-center-to-profit-driver`: "How a 12-person team
  created $5M in value", $2.8M annual cost.
- `when-your-startup-needs-emergency-cto-leadership`: three invented crisis
  narratives opening "let me share three situations".

Left deliberately: the Weak/Strong Example pair in `internal-product-teams` reads
as a claim out of context but is a labelled template teaching the reader to
phrase their OWN success story. Deleting it would have been pattern-matching
rather than reading.

Ratchet re-proved exact at the new count - dropped to 8, failed with "Expected 9
to be <= 8", restored.

FLAGGED, not fixed - a marker candidate rather than a hand-sweep:
`tdd-workflow-automation-rails-teams` §"Measuring TDD Automation Impact" is ~103
lines of invented before/after metrics, survey results and a "$562,500 annual
value" ROI, presented as measured. It is not a ratchet hit and I found it
incidentally, so hand-fixing it here would be scope creep on a zero-traffic page.
The right fix is a marker for the fabricated-metrics-table shape, which would
catch it and its siblings at once.

Gates: marketing_copy_test 4 runs / 10 assertions / 0 failures. `bin/hugo-build`
green. Content + test only, so the visual suites do not apply.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_016PUkwFTsiv7EB2DYKogbpg