Diff

baseline p9-46-torque-baseline · d88f256424e0 against candidate p9-48-torque-one-bolt-low · 72e97206b080

Both runs' records agree. Every check behind this is on Proof

The short answer

The runs first differ at read_tool_log, where input tool_log.csv changed.

No stage comes before it. Both stages after it differ.

  1. read_tool_logFirst difference
  2. torque_checkChanged after it
  3. station_decisionChanged after it

Checked from the records. The same answer every time.

What may have happened

AI-generated, may be wrong · model anthropic:claude-sonnet-4-6

Hide the detailsShow the details
Lead (possible culprit)

The input file `tool_log.csv` differs between runs and is the first difference detected; this is the most likely source of the divergence.

Note

The candidate `tool_log.csv` is the same size (`61` bytes) as the baseline but has a different digest, suggesting a changed torque value rather than a structural change.

Note

The final `decision.json` differs in size (`39` bytes baseline vs `41` bytes candidate), confirming the changed input propagates to the pipeline output.

Checks

onetrace can run these. A lead is confirmed or ruled out only by what they report. This demo runs none of them.

  • onetrace diff a b --text --stage read_tool_log
  • onetrace diff a b --text --stage torque_check
  • onetrace diff a b --text --stage station_decision
6 claims checked against the record · 3 left out
  • 3 quotes a value its cites don't hold

Its record: explain/2026-10-25T13-42-26.188702Z/, onetrace-verify: 88 pass, 0 fail, 3 not-run -> PASS (exit 0)

What changed between the two runs

The two runs differ.

First difference: stage 1 · read_tool_log.

No stage comes before it.

input tool_log.csv changed: 8313fc7e9e28 -> b0453a3fe0c8

Both runs' records agree. No recorded setting of read_tool_log differs. Its output differs, and all 2 stages after it changed too.

onetrace diff runs/p9-46-torque-baseline runs/p9-48-torque-one-bolt-low --text

recorded as a straight line (no graph recorded)

read_tool_log · First differenceread_tool_logFirst differencetorque_check · Changed after ittorque_checkChanged after itstation_decision · Changed after itstation_decisionChanged after it
  1. read_tool_logFirst difference
  2. torque_checkChanged after it
  3. station_decisionChanged after it

A link shows which recorded output a stage read. It does not by itself prove that one stage caused the next.

Settings that changed

No recorded setting differs between the two runs.

The first difference is at stage 1 · read_tool_log. No smaller set was reported.

The text diff

Shown with at most 200 lines per output (the SDK's default).

Open the text diff on its own page

first difference: read_tool_log — output text changed; settings unchanged; input tool_log.csv changed

read_tool_log · First difference

settings: none changed

input tool_log.csv (externally-sourced) changed: 8313fc7e9e28 -> b0453a3fe0c8; content not stored

torques.json application/json Baseline: verified Candidate: verified

1 line removed · 1 added

22    "torque_nm": [
33      "118.4",
44      "121.0",
5-     "119.6",
5+     "104.2",
66      "122.3",
77      "120.8"
88    ]

torque_check · Changed after it

settings: none changed

input torques.json (operator-authored) changed: d8b870178f9c -> 5b52c509e579; content not stored

check.json application/json Baseline: verified Candidate: verified

1 line removed · 6 added

11  {
2-   "outside": [],
2+   "outside": [
3+     {
4+       "bolt": "3",
5+       "torque_nm": "104.2"
6+     }
7+   ],
38    "window_nm": "110.0 to 130.0"
49  }

station_decision · Changed after it

settings: none changed

input check.json (operator-authored) changed: 517dfe129ffd -> 62efd69227d4; content not stored

decision.json application/json Baseline: verified Candidate: verified

2 lines removed · 2 added

11  {
2-   "bolts_outside": "0",
2+   "bolts_outside": "1",
3-   "decision": "pass"
3+   "decision": "rework"
44  }

Command: onetrace diff runs/p9-46-torque-baseline runs/p9-48-torque-one-bolt-low --text

Repeat this yourself:

onetrace diff runs/p9-46-torque-baseline runs/p9-48-torque-one-bolt-low

Other runs that recorded the same value

Matching rule: the same input tool_log.csv digest at stage 1 · read_tool_log, by exact equality:
sha256:b0453a3fe0c8….sha256:b0453a3fe0c86e73e4d16897f1a78a1caf85cbde410860ee5cbe8e4138c4878e

1 of 3 stored runs in P9 recorded the same input tool_log.csv digest at stage 1 · read_tool_log, the candidate among them.

Nothing here says what the change did in those runs.

This record shows what was verified and evaluated at the time — not what was ultimately true.