feat(blog): add work-email-finding-bench receipt post

[JayZeeDesign]

What this does

Adds /blog/work-email-finding-bench, a receipt post comparing work-email finding across vendors (treg, Clay, Monid, Freckle, Deepline) based on a measured 2026-09-16 neutral run.

Key findings from the 292-person bench:

• Quality tie: treg, Clay, Freckle at 262-264 exact matches (difference is noise)
• Cost gap: treg $0.0056/correct vs Clay $0.0395/correct (~7x)
• Latency: treg 0.44s median (per-call) vs Clay 11.2s

Includes:

• Full data table with all measured numbers (find rates, exact match, precision, cost)
• Fair Clay framing ("when to choose Clay" section for visual/GTM/spreadsheet workflows)
• Disclosures: list bias (all had published emails), MillionVerifier re-verify, 3 rows 502 bug
• Links to /pricing, /people-search, /workflows/find-and-verify-a-lead-list, /use-cases/lead-enrichment-for-ai-agents
• Blog index card, sitemap entry, route ownership

Source data: 2026-09-16 neutral run. Row-level CSV files stay private.

How it was tested

uv run --frozen python -m pytest tests/test_blog.py -v  # 15 tests pass
uv run --frozen python -m pytest tests/test_seo.py -v   # 69 tests pass (includes sitemap 200 check)
uv run --frozen python -m pytest tests/test_app_roles.py -v  # 7 tests pass (route snapshot)
uv run lint-imports  # 14 contracts kept

New tests in tests/test_blog.py:

• 200 status, canonical, H1, breadcrumbs
• Key measured numbers present (90.4%, $0.0056, $0.0395, etc.)
• Links to related pages
• Disclosures present (list bias, MillionVerifier, 502 bug)
• Fair Clay framing present
• No em-dashes (product constraint)

Checklist

• uv run --with pytest-xdist pytest -n auto -q passes locally
• Added or updated tests for the change (if it affects behavior)
• Updated the relevant docs/context/ fragment (if a subsystem changed) — N/A, blog post only
• No secrets in the diff (keys, tokens, .env values)

  

[cursor]

Folded into #551.