Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Run the agent-assisted Smithers upgrade workflow: fetch changelogs, upgrade with a cheap agent, and escalate to a smart agent only when needed. Run `smithers upgrade --help` for usage details.
.claude/skills/smithersai-smithers-upgrade/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | -79% | 0% |
| case-02 | ✗→✓ | ▲ Improved | -58% | 0% |
| case-03 | ✗→✓ | ▲ Improved | -54% | 0% |
| case-05 | ✗→✓ | ▲ Improved | -61% | 0% |
| case-06 | ✗→✓ | ▲ Improved | -16% | 0% |
Run the agent-assisted Smithers upgrade workflow: fetch changelogs, upgrade with a cheap agent, and escalate to a smart agent only when needed.
| Flag | Type | Default | Description | |------|------|---------|-------------| | --interactive | boolean | false | Force the full-screen interactive TUI monitor (TTY only). | | --detach | boolean | false | Launch the upgrade workflow in the background and print the run ID. | | --dryRun | boolean | false | Fetch changelogs and plan the upgrade without changing the install. | | --runId | string | | Explicit run ID for the upgrade workflow. | | --root | string | | Tool sandbox root directory. | | --logDir | string | | NDJSON event logs directory. | | --backend | string | | Storage backend for the upgrade workflow run. | | --authToken | string | | Bearer token passed to the interactive monitor gateway client. |
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 12,482 | 2,121 | -83% | 1 | 1 | 0% | 1,979 | 409 | -79% | 0 | 0 | — |
case-02 | fail→pass | 7,042 | 1,617 | -77% | 1 | 1 | 0% | 1,109 | 462 | -58% | 0 | 0 | — |
case-03 | fail→pass | 6,339 | 1,856 | -71% | 1 | 1 | 0% | 1,151 | 525 | -54% | 0 | 0 | — |
case-04 | pass→pass | 4,432 | 1,516 | -66% | 1 | 1 | 0% | 795 | 455 | -43% | 0 | 0 | — |
case-05 | fail→pass | 8,863 | 2,509 | -72% | 1 | 1 | 0% | 1,514 | 598 | -61% | 0 | 0 | — |
case-06 | fail→pass | 13,869 | 1,713 | -88% | 1 | 1 | 0% | 591 | 497 | -16% | 0 | 0 | — |
case-11 | fail→pass | 6,675 | 1,702 | -75% | 1 | 1 | 0% | 1,191 | 553 | -54% | 0 | 0 | — |
case-07 | fail→pass | 3,805 | 4,386 | +15% | 1 | 1 | 0% | 504 | 474 | -6% | 0 | 0 | — |
case-08 | fail→pass | 13,199 | 1,668 | -87% | 1 | 1 | 0% | 2,982 | 508 | -83% | 0 | 0 | — |
case-09 | fail→pass | 6,234 | 2,248 | -64% | 1 | 1 | 0% | 1,015 | 567 | -44% | 0 | 0 | — |
case-10 | fail→pass | 6,531 | 1,571 | -76% | 1 | 1 | 0% | 1,114 | 497 | -55% | 0 | 0 | — |
case-12 | fail→pass | 10,989 | 1,796 | -84% | 1 | 1 | 0% | 1,574 | 495 | -69% | 0 | 0 | — |
case-13 | fail→pass | 5,761 | 1,916 | -67% | 1 | 1 | 0% | 979 | 563 | -42% | 0 | 0 | — |
case-14 | pass→pass | 12,064 | 2,092 | -83% | 1 | 1 | 0% | 2,114 | 540 | -74% | 0 | 0 | — |
case-15 | fail→pass | 30,588 | 1,231 | -96% | 1 | 1 | 0% | 2,720 | 390 | -86% | 0 | 0 | — |
case-16 | fail→pass | 9,399 | 2,231 | -76% | 1 | 1 | 0% | 1,594 | 616 | -61% | 0 | 0 | — |
case-17 | fail→pass | 12,961 | 1,865 | -86% | 1 | 1 | 0% | 2,066 | 526 | -75% | 0 | 0 | — |
case-18 | fail→pass | 12,006 | 1,752 | -85% | 1 | 1 | 0% | 1,959 | 540 | -72% | 0 | 0 | — |
case-19 | fail→pass | 14,284 | 2,332 | -84% | 1 | 1 | 0% | 2,387 | 558 | -77% | 0 | 0 | — |
case-20 | fail→pass | 7,958 | 1,528 | -81% | 1 | 1 | 0% | 1,159 | 444 | -62% | 0 | 0 | — |
case-21 | pass→pass | 12,179 | 3,392 | -72% | 1 | 1 | 0% | 2,151 | 739 | -66% | 0 | 0 | — |
case-22 | pass→pass | 9,633 | 9,254 | -4% | 1 | 1 | 0% | 1,807 | 2,013 | +11% | 0 | 0 | — |
case-23 | pass→pass | 12,021 | 10,591 | -12% | 1 | 1 | 0% | 1,874 | 1,848 | -1% | 0 | 0 | — |
case-24 | pass→pass | 8,013 | 5,591 | -30% | 1 | 1 | 0% | 1,450 | 1,396 | -4% | 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. 24 cases were attempted, and 23 counted toward the lift figure. The other 1 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. The headline lift of +75 percentage points is the difference between those two pass rates over the 23 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.