Found during the new-developer first-run journey audit, #10264. Low practical impact, disproportionate trust impact.
Both numbers re-verified by the PM against origin/main
Claim 1 — README.md:143-150. The text asserts examples/app-crm is "31 files, 1,792 lines, roughly 16k tokens", then prints its own verification command and says "Count it yourself:"
find examples/app-crm/src -name '*.ts' -not -name '*.test.ts'| xargs cat | wc -l
Running that exact command on origin/main:
| README | Actual |
|---|
| Files | 31 | 31 ✓ |
| Lines | 1,792 | 1,930 ✗ (+138) |
Claim 2 — README.md:265. The <details> summary reads "72 published packages". The table it expands to contains ~45 linked package rows (grep -c '^| \[' on the section). For reference, the repo has 69 non-private package.json files under packages/. Whichever number is intended, the summary and the table it introduces do not agree with each other.
Why this is worth a card despite being cosmetic
The line-count miss is the one that stings, because the README does not merely assert the number — it hands the reader the command and invites them to check. A newcomer who accepts that invitation gets a different answer. That converts a confidence-building gesture into the first thing on the page that makes them doubt everything else on it — including the tests 6,507 passing badge two screens up, which no outside reader can verify at all.
A page that says "count it yourself" and is wrong when you do teaches the reader that its numbers are decorative.
Note for whoever takes this
Both numbers are the kind that drift silently with every merge. Consider whether either belongs under a gate rather than being re-fixed by hand.
Filed by the PM loop from the audit in #10264. Not yet graded or routed.
Found during the new-developer first-run journey audit, #10264. Low practical impact, disproportionate trust impact.
Both numbers re-verified by the PM against
origin/mainClaim 1 —
README.md:143-150. The text assertsexamples/app-crmis "31 files, 1,792 lines, roughly 16k tokens", then prints its own verification command and says "Count it yourself:"Running that exact command on
origin/main:Claim 2 —
README.md:265. The<details>summary reads "72 published packages". The table it expands to contains ~45 linked package rows (grep -c '^| \['on the section). For reference, the repo has 69 non-privatepackage.jsonfiles underpackages/. Whichever number is intended, the summary and the table it introduces do not agree with each other.Why this is worth a card despite being cosmetic
The line-count miss is the one that stings, because the README does not merely assert the number — it hands the reader the command and invites them to check. A newcomer who accepts that invitation gets a different answer. That converts a confidence-building gesture into the first thing on the page that makes them doubt everything else on it — including the
tests 6,507 passingbadge two screens up, which no outside reader can verify at all.A page that says "count it yourself" and is wrong when you do teaches the reader that its numbers are decorative.
Note for whoever takes this
Both numbers are the kind that drift silently with every merge. Consider whether either belongs under a gate rather than being re-fixed by hand.
Filed by the PM loop from the audit in #10264. Not yet graded or routed.