# Evidence: every claim, read back

Standing of each row of `docs/CLAIMS.md` at revision `a08b62374a15e107c43d5eded56af979a3ea495c`, from `python3 evals/claims.py --json`.

| Standing | Claim | Reason |
|---|---|---|
| holds | A recorded result is answered `holds`, `fails`, `stale` or `unknown`, with a reason; `unknown` is never turned into another answer | the code and the test it names exist |
| holds | A passed check holds for the bytes it tested and stops holding when one changes, however recent it is | the code and the test it names exist |
| holds | The answer names the file that changed | the code and the test it names exist |
| holds | No route moves a task to review or done unless every check holds, or a waiver with a reason is recorded and shown | the code and the test it names exist |
| holds | A result is shared by every task whose files are the same bytes, and is executed once | the code and the test it names exist |
| holds | A change to a file no domain claims is `unknown`; a check naming an undeclared domain is `unknown` | the code and the test it names exist |
| holds | A declared domain holds what its patterns match on disk, so an input git ignores is part of what a check stood on | the code and the test it names exist |
| holds | A check marked isolated can read only the files its result is filed under, and what it writes is discarded | the code and the test it names exist |
| holds | Only a person accepts a result, and the acceptance is of one version | the code and the test it names exist |
| holds | A work product is current while its declared inputs are the same bytes; with none declared it is `unknown` | the code and the test it names exist |
| holds | A new or lapsed instruction makes work closed under the old ones `stale` | the code and the test it names exist |
| holds | The instruction that governed a scope on a past day is still answerable after it was replaced | the code and the test it names exist |
| holds | A replacement governs from its own date, inside its own scope, until it lapses | the code and the test it names exist |
| holds | What governs a task and what was learned reach the agent at start, in the original words | the code and the test it names exist |
| holds | A reminder shown is an exposure; only a judged outcome is a repeat | the code and the test it names exist |
| holds | A setup is credited only when it was observed to run | the code and the test it names exist |
| holds | A lapsed lease is over | the code and the test it names exist |
| holds | A refused start takes nothing | the code and the test it names exist |
| holds | A label is read strictly and a missing one is counted apart | the code and the test it names exist |
| holds | A decision-bank run keeps the answers it scored and their hash, so the score can be recomputed | the code and the test it names exist |
| holds | A change is not confirmed by a number with no source, and a prediction is written once | the code and the test it names exist |
| holds | A number quoted here is stale once the code it was measured on changes, and the answer names the file | the code and the test it names exist |
| holds | An isolated check sees a fixed environment, not the caller's | the code and the test it names exist |
| holds | Each isolated result records what was closed, found by trying; where the network and the rest of the workspace can be closed they are, and a strict check does not run where they cannot | the code and the test it names exist |
| holds | The setup answers can say which files each check reads | the code and the test it names exist |
| holds | No row of this file names code or a test that is not there, or a number its recorded run does not contain | the code and the test it names exist |
| holds | The demo runs from a clean copy and prints each idea with real output | the code and the test it names exist |
| holds | Earlier releases had the counterexample cases: every one at the first, five at the second | measured on these bytes |
| holds | The tests notice when a protection is removed. The first run, with 16 faults, caught 15; the missing test was added | measured on these bytes |
| holds | Reuse by basis relied on no stale result and executed what an oracle would | measured on these bytes |
| holds | Age limits trade reruns against stale reuse | measured on these bytes |
| holds | In the open, under-declared inputs are reused wrongly | measured on these bytes |
| holds | Isolated, a check that reads what it did not declare never passes, and correct declarations cost no extra executions. Predicted before the code was written | measured on these bytes |
| holds | Asking costs time that grows with the tree | measured on these bytes |
| holds | Isolating a check costs milliseconds. Measured in pairs that differ only in the isolation, order alternating | measured on these bytes |
| holds | The suite passes on Python 3.9 and 3.12, Linux and macOS | counted |
| holds | README.md repeats: 151 tests | the recorded run says the same |
| holds | README.md repeats: 22 of 22 fail at `071a348`, 5 of 22 fail at `89f36ce`, 22 of 22 pass here | the recorded run says the same |
| holds | README.md repeats: 20 of 20 removals are caught | the recorded run says the same |
| holds | README.md repeats: 60 executions and 0 wrong reuses | the recorded run says the same |
| holds | README.md repeats: 150 wrong reuses in 360 | the recorded run says the same |
| holds | README.md repeats: 0 of 366 runs passed | the recorded run says the same |

Full text with code, tests and what is not claimed: https://github.com/JustinBen0131/ThesisAnalysisOS/blob/a08b62374a15e107c43d5eded56af979a3ea495c/docs/CLAIMS.md
