Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Fork from a checkpoint and resume execution (time travel). Run `smithers replay --help` for usage details.
.claude/skills/smithersai-smithers-replay/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | -58% | 0% |
| case-02 | ✗→✓ | ▲ Improved | -49% | 0% |
| case-04 | ✗→✓ | ▲ Improved | -39% | 0% |
| case-05 | ✗→✓ | ▲ Improved | -40% | 0% |
| case-06 | ✗→✓ | ▲ Improved | -30% | 0% |
Fork from a checkpoint and resume execution (time travel).
| Name | Type | Required | Description | |------|------|----------|-------------| | workflow | string | yes | Workflow ID (from smithers workflow list) or path to a .tsx workflow file |
| Flag | Type | Default | Description | |------|------|---------|-------------| | --runId | string | | Source run ID to replay from | | --frame | number | | Frame number to fork from | | --node | string | | Node ID to reset to pending | | --input | string | | Input overrides as JSON string | | --label | string | | Branch label for the fork | | --restoreVcs | boolean | false | Restore jj filesystem state to the source frame's revision |
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-14 | pass→pass | 6,228 | 1,896 | -70% | 1 | 1 | 0% | 1,162 | 516 | -56% | 0 | 0 | — |
case-15 | pass→pass | 10,615 | 2,300 | -78% | 1 | 1 | 0% | 1,675 | 617 | -63% | 0 | 0 | — |
case-01 | fail→pass | 9,428 | 2,792 | -70% | 1 | 1 | 0% | 1,582 | 657 | -58% | 0 | 0 | — |
case-02 | fail→pass | 11,710 | 2,950 | -75% | 1 | 1 | 0% | 1,401 | 717 | -49% | 0 | 0 | — |
case-03 | fail→fail | 7,552 | 2,556 | -66% | 1 | 1 | 0% | 1,269 | 644 | -49% | 0 | 0 | — |
case-04 | fail→pass | 6,739 | 2,946 | -56% | 1 | 1 | 0% | 1,261 | 771 | -39% | 0 | 0 | — |
case-05 | fail→pass | 8,019 | 3,353 | -58% | 1 | 1 | 0% | 1,483 | 889 | -40% | 0 | 0 | — |
case-06 | fail→pass | 6,598 | 2,506 | -62% | 1 | 1 | 0% | 1,018 | 709 | -30% | 0 | 0 | — |
case-07 | fail→pass | 5,082 | 2,539 | -50% | 1 | 1 | 0% | 1,099 | 663 | -40% | 0 | 0 | — |
case-08 | pass→pass | 13,010 | 1,728 | -87% | 1 | 1 | 0% | 2,126 | 556 | -74% | 0 | 0 | — |
case-09 | fail→pass | 7,976 | 2,416 | -70% | 1 | 1 | 0% | 1,316 | 664 | -50% | 0 | 0 | — |
case-10 | fail→pass | 8,278 | 2,346 | -72% | 1 | 1 | 0% | 1,312 | 553 | -58% | 0 | 0 | — |
case-11 | fail→pass | 6,806 | 2,940 | -57% | 1 | 1 | 0% | 1,291 | 635 | -51% | 0 | 0 | — |
case-12 | fail→pass | 9,560 | 2,645 | -72% | 1 | 1 | 0% | 1,529 | 630 | -59% | 0 | 0 | — |
case-13 | pass→pass | 11,889 | 1,877 | -84% | 1 | 1 | 0% | 1,009 | 495 | -51% | 0 | 0 | — |
case-16 | fail→pass | 7,317 | 2,912 | -60% | 1 | 1 | 0% | 1,165 | 726 | -38% | 0 | 0 | — |
case-17 | pass→pass | 4,769 | 1,966 | -59% | 1 | 1 | 0% | 778 | 516 | -34% | 0 | 0 | — |
case-18 | fail→pass | 4,572 | 1,830 | -60% | 1 | 1 | 0% | 838 | 480 | -43% | 0 | 0 | — |
case-19 | pass→pass | 7,683 | 3,374 | -56% | 1 | 1 | 0% | 1,256 | 662 | -47% | 0 | 0 | — |
case-20 | fail→pass | 7,781 | 2,077 | -73% | 1 | 1 | 0% | 1,284 | 507 | -61% | 0 | 0 | — |
case-21 | pass→pass | 7,562 | 3,051 | -60% | 1 | 1 | 0% | 1,264 | 733 | -42% | 0 | 0 | — |
case-22 | fail→pass | 12,709 | 6,172 | -51% | 1 | 1 | 0% | 2,137 | 1,268 | -41% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted. The headline lift of +64 percentage points is the difference between those two pass rates over the 22 comparable cases.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.