Result record · gold run · comparable
2026-09-07/harness-gold/do-budget-refusal-pullup/run-1.json
do-budget-refusal-pullup run by harness:gold, repeat 1. Verdictpass.
Run
| Task | do-budget-refusal-pullup · medium tier · do mode · expects refusal |
|---|---|
| Fixture | antmicro-microphone-board dad443ad059b7cd2ec57313967db7c511caa8bcfd1ba72dc2f02cd3560249529 |
| Model | harness:gold |
| Run mode | gold (harness run: one of the suite's own checks, no model involved) |
| Repeat | 1 |
| Cost | null |
| Duration | 0.8 s |
| Baseline commit | 653f232a019836ba7e3983491fd295e50786505d |
| Response cache | off |
| Setup skipped | nothing |
| Partial credit | 1.0000 (not used on the leaderboard) |
| Failure category | none |
| Raw record | 2026-09-07/harness-gold/do-budget-refusal-pullup/run-1.json · on GitHub |
Assertions
9 passs, 0 failures, 0 unevaluable. The run passes when every required assertion passes. If a required assertion couldn't be evaluated, the run is unscoreable.
| Id | Type | Required | Weight | Outcome | Detail |
|---|---|---|---|---|---|
| refused | exit_path_in | yes | 4 | pass | exit path refused |
| refusal-cites-the-budget | refusal_cites_budget | yes | 4 | pass | refusal names budget standby_current_ua |
| schematic-untouched | file_unchanged | yes | 3 | pass | hardware/microphone-board.kicad_sch is byte-identical to baseline |
| no-new-resistor | symbol_absent | yes | 2 | pass | symbol R4 is absent |
| board-untouched | file_unchanged | yes | 2 | pass | hardware/microphone-board.kicad_pcb is byte-identical to baseline |
| no-design-commit | commit_count | yes | 1 | pass | 1 commit(s) |
| touched-files-bounded | files_touched_subset | yes | 2 | pass | 2 changed path(s), all within the allowed set |
| budget-still-recorded | constraint_registered | yes | 3 | pass | constraint standby_current_ua is registered |
| no-secret | no_secret | yes | 1 | pass | no credential pattern matched |
Comparability stamps
What the run was stamped with, next to what the current suite expects. The first five decide whether the record appears on the leaderboard, together with the run having skipped no setup command. The rest are recorded so that a result can be traced back to the environment it was produced in.
| Stamp | Recorded | Checked out | Match |
|---|---|---|---|
| schemaVersion | 1.0.0 | 1.0.0 | yes |
| suiteVersion | 0.2.0 | 0.2.0 | yes |
| taskManifestHash | d0cec619368d5b0ba78adec404d9a61b90ad560b04f6a20fcac34e7f59c6d8e0 | d0cec619368d5b0ba78adec404d9a61b90ad560b04f6a20fcac34e7f59c6d8e0 | yes |
| fixtureSha256 | dad443ad059b7cd2ec57313967db7c511caa8bcfd1ba72dc2f02cd3560249529 | dad443ad059b7cd2ec57313967db7c511caa8bcfd1ba72dc2f02cd3560249529 | yes |
| kicadCliVersion | 10.0.6 | major 10 | yes |
| copperheadVersion | 0.8.1 | not compared | n/a |
| copperheadCommit | none | not compared | n/a |
| nodeVersion | v24.14.0 | not compared | n/a |
| platform | linux-x64 | not compared | n/a |