Get ThruLine

Specimen 02

The run was one version older.

The tests passed, and the agent said so. They passed on 9f2ad3e — one version (one exact commit of the code) before e71c5b0, the one you are approving.

Recorded on ThruLine’s checkout-web demo record. No customer data. [swap in a corpus chain before launch]

The chain

Six questions, asked in order; each answer is a line from the record. Proof holds through ② and stops at ③: the run the claim rests on was made on an older version.

Claude Code said “Checkout total fixed.” “Tests pass.” · step 52 · demo record

Before the re-run

  1. ① Said

    What did the agent claim?

    "Tests pass." Claude Code · 16:03 step 52

  2. ② Rests on

    What is that claim based on?

    4 passed, 0 failed step 46 cited by the agent

  3. ③ Ran on

    Which version did that actually run on?

    version 9f2ad3e 1 version older

  4. ④ Since

    Did anything it depends on change after it?

    checkout/api.py changed · step 56 after the run

  5. ⑤ Re-checked

    Has anything checked it since?

    nothing on e71c5b0 with checkout/api.py in scope

  6. ⑥ So
    Not proven

    So what do we actually know?

    version e71c5b0 the claim may be entirely true.

③ Ran on · where proof stops

Which version did that actually run on?

On 9f2ad3e. The run the agent cited, step 46, passed on the code as it stood before the edit at step 56. The pass is real, and it belongs to 9f2ad3e. You are approving e71c5b0.

  1. 4b19a0c run 26
  2. 9f2ad3e run 46, needs a re-check
  3. e71c5b0 · now edit · step 56
Read it left to right, oldest first. Run 46 sits left of the edit at step 56, so its pass is proof for an older version: it needs a re-check before it can count for e71c5b0.

Proof stops here

Nothing on e71c5b0 has run with checkout/api.py in scope.

What was looked at · 2 finished runs, 1 step with no result
Step Ran on Result Why it doesn’t count here
step 26 4b19a0c 1 failed, 3 passed An older version, and it failed.
step 46 9f2ad3e 4 passed, 0 failed The run the agent cited. It passed on 9f2ad3e, one version older.
step 60 No result yet A step with no result never counts.
Check it again

pytest -q tests/test_totals.py checkout/ on e71c5b0 · asks once for this project

The re-run

One press of Check it again runs the check on the version you are about to approve. It asks once per project, installs nothing and does not touch the network. In v0.2.0 the button runs cargo test; a pytest check like this one is re-run from a terminal with thruline check -- pytest ….

thruline ui · Change · Fix double-applied discount on checkout total After the re-run

1 · You press

Check it again

pytest -q tests/test_totals.py checkout/ on e71c5b0 · asks once for this project

2 · It runs

Passed

step 62 · exit 0 · on e71c5b0 3 passed, 0 failed

3 · ⑥ So

Proven · version e71c5b0 · tests/test_totals.py, checkout/

“Tests pass.” now rests on a run on this version.

  1. ① Said
  2. ② Rests on
  3. ③ Ran on
  4. ④ Since
  5. ⑤ Re-checked

    re-run · step 62 exit 0

  6. ⑥ So
    Proven

    version e71c5b0

What to notice: step 46 still ran on 9f2ad3e — the record never changes. The proof is a new run, step 62, on e71c5b0.

The receipt

A receipt is the record of exactly what ran: the command, the version it ran on, the exit code and the counts, with a sha256 fingerprint of the whole. “Proven” always points at one — a recorded run, not a re-read.

Receipt · step 62 · exit 0 · 3 passed, 0 failed · version e71c5b0

command
pytest -q tests/test_totals.py checkout/
scope
tests/test_totals.py, checkout/
outcome
Passed
record
checkout-web · demo record
sha256
[receipt sha256]

This is the checkout-web demo record, so its hash stays a placeholder. A real signed receipt, from ThruLine’s qualification run on a synthetic test project, is on the home page.

What this shows

A pass belongs to the version it ran on. Step 46 is a real pass — for 9f2ad3e. Until a run on e71c5b0 says the same, ThruLine says “Not proven”, and the claim may be entirely true.