A claim names the test that settles it, a recorded run is judged against the commit it ran at, and a claim with no recorded run says so.
Claims outrun proof. The documentation says one thing, the runtime does another, and nobody can show which is true.
What shows that a statement about the product is true: claims joined to the executions recorded for their tests, use cases the tool runs against itself, measurements of speed and of token cost that are recorded rather than typed, and the provenance of every record. A run against this commit is never reported as the same thing as a run whose inputs merely have not changed.
Not in this domain: the verdict on one task, which is completion; and the rules a claim is about, which is governance.
Each feature names this domain in its own file; the domain lists none of them. A feature added with this domain is on this page at the next build, and one removed is gone.
What a coding session consumes with Majordomus and without it, from matched runs judged by the same hidden tests, with every number labelled by how it was obtained and published only past a declared rule.
2 of 3 claims guaranteed with a named test
Who did what, under which task, at which commit, in which session is a ledger the tool writes and a person can read back, not a memory of a conversation.
6 of 9 claims guaranteed with a named test · 2 use cases
What a person does with the tool is one file with a scenario the tool runs against itself; the page shows that execution, and a public command no use case runs is a gap the tool reports.
4 of 4 claims guaranteed with a named test · 1 use case
Every claim is joined to the execution recorded for its test — the outcome, the commit, the tree, the digest and the command that produces it again — and a run against this commit is never reported as the same thing as a run whose inputs merely have not changed.
12 of 12 claims guaranteed with a named test
Every externally callable operation is timed directly, over MCP and over HTTP, the denominator is generated from the registry, and a regression against the accepted baseline fails the check.
6 of 6 claims guaranteed with a named test · 1 use case
Operational moments from the Why catalogue that name a rule, a command, a capability or a claim of a feature in this domain. Nothing here was assigned.
Every claim a feature of this domain names, once, with the status it declares — guaranteed, advisory, planned or rejected — and the test that settles it. Whether a recorded run supports a claim is its own page's verdict.
Scenarios the tool runs against itself every time this site is built.