Where: README.md's "Show all 61 explainers" collapsible table, and CONTRIBUTING.md's explainer list (used by contributors to check for duplicates before starting a new one, per the issue template's own instructions).
The gap: both tables are missing 5 of the 61 real, published explainer files - all correctly built into the site (present in assets/explainers-data.json, sitemap.xml, and llms-full.txt), just never added to these two tables:
import re, glob
text = open('README.md').read()
start = text.find('## Explainers'); end = text.find('## ', start+10)
section = text[start:end]
linked = set(re.findall(r'\(explainers/([a-z0-9-]+)\.md\)', section))
real = set(f.split('/')[-1].replace('.md','') for f in glob.glob('explainers/*.md'))
print(sorted(real - linked))
# ['conditional-demographic-parity', 'differential-privacy', 'reject-option-classification', 'simpsons-paradox', 'subgroup-fairness']
grep -c confirms 56 table rows in README's section against 61 real files; grep finds zero mentions of any of these 5 slugs in CONTRIBUTING.md.
Why this matters beyond an incomplete list: scripts/check_explainer_count.py already tracks the count mention ("Show all 61 explainers") and confirms it's correct - the count was bumped correctly when these 5 shipped (via #541/#562/#565), it's just the actual table rows that were never added. The new-explainer issue template tells contributors to check this table for duplicates before starting - a contributor doing exactly that right now would not see these 5 topics as already covered.
Fix direction: add a row (README.md) and a mention (CONTRIBUTING.md) for each of the 5 missing explainers, matching the existing format. Consider whether check_explainer_count.py (or a new check) should also verify every real explainer has a table row, not just that the count number is right - this exact "count matches but the underlying list drifted" gap has now happened at least once.
Where:
README.md's "Show all 61 explainers" collapsible table, andCONTRIBUTING.md's explainer list (used by contributors to check for duplicates before starting a new one, per the issue template's own instructions).The gap: both tables are missing 5 of the 61 real, published explainer files - all correctly built into the site (present in
assets/explainers-data.json,sitemap.xml, andllms-full.txt), just never added to these two tables:grep -cconfirms 56 table rows in README's section against 61 real files;grepfinds zero mentions of any of these 5 slugs inCONTRIBUTING.md.Why this matters beyond an incomplete list:
scripts/check_explainer_count.pyalready tracks the count mention ("Show all 61 explainers") and confirms it's correct - the count was bumped correctly when these 5 shipped (via #541/#562/#565), it's just the actual table rows that were never added. The new-explainer issue template tells contributors to check this table for duplicates before starting - a contributor doing exactly that right now would not see these 5 topics as already covered.Fix direction: add a row (README.md) and a mention (CONTRIBUTING.md) for each of the 5 missing explainers, matching the existing format. Consider whether
check_explainer_count.py(or a new check) should also verify every real explainer has a table row, not just that the count number is right - this exact "count matches but the underlying list drifted" gap has now happened at least once.