ProofOps · measured use
The numbers, from the people who opted in
Every figure here arrived as an explicit opt-in: the "Send anonymous diagnostics"
checkbox offered once, after the first run. Saying yes sends nine counters; saying no is remembered
and never asked again. Installs that never opt in are simply not counted,
which is the honest denominator: this page reports its own coverage, not the internet's.
Twenty seconds, no account, nothing leaves your machine
pipx install git+https://github.com/rasrobo/proofops && proofops demo
[PASS] reachable: the endpoint answers 200
[FAIL] says-ok: the service says it is healthy
observed: {"expected": "ok", "observed": "degraded"}
What is being verified
By check type
| check type | passed | failed | not run |
|---|
Versions in use
Platforms
What we do with these diagnostics. They are the only feedback this tool gets. The check type
with the most checks that
could not run is the next thing we fix, and the type with the
most failures gets the next diagnostic. Opting in is a vote for what gets built next.
What this page proves. That this many installs opted in and reported these cumulative
counts. It does not prove the counts are true, that the work verified was important, or that a
single defect was caught — a check that passes is not a defect avoided.
What it can never contain. No customer names, no repository or file paths, no hostnames,
no plan names, no receipt contents, no command output, no email addresses. The payload is six
fields with an allow-listed version and platform; the sanitizer that enforces it is tested, and
the server refuses anything else.
Loading…