Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →UI test recipe -- composes browser-record (capture) + browser-replay (verify) so every test produces a replayable RVF artifact, not an ephemeral run
.claude/skills/ruvnet-browser-test/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | -32% | 0% |
| case-03 | ✗→✓ | ▲ Improved | 492% | 0% |
| case-09 | ✗→✓ | ▲ Improved | -24% | 0% |
| case-10 | ✗→✓ | ▲ Improved | -48% | 0% |
| case-15 | ✗→✓ | ▲ Improved | -52% | 0% |
Automated UI testing. Now backed by a recorded RVF session container instead of an ephemeral run, so every test produces a replayable artifact.
browser-record:--kind browser-session.browser_open, browser_click, browser_fill, browser_type, browser_select. Each action emits a trajectory-step.browser_wait before assertions.browser_get-text / browser_get-value / browser_get-title / browser_get-url. Validation outcomes go into findings.md inside the RVF container.<step-id>.png.trajectory-end --verdict pass|fail, rvf compact, AgentDB index in browser-sessions.--against <prior-session-id>: invoke browser-screenshot-diff to compare the new run with a baseline.browser_back / browser_forward for history navigationbrowser_reload to refresh the pagebrowser_scroll to scroll to elements or coordinatesbrowser_close alone — it ends with the session-end protocol.browser-selectors (host:intent), so the next test can find them by embedding similarity.aidefence_is_safe before any LLM-facing summary; injection-flagged content is quarantined to findings.md./ruflo-browser replay can re-drive.browser_wait before assertions to handle async rendering.--against <id> on the next run.browser_eval for custom JavaScript assertions — but redact any returned strings via the aidefence_is_safe gate before logging.| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 22,610 | 10,757 | -52% | 1 | 1 | 0% | 4,219 | 2,863 | -32% | 0 | 0 | — |
case-02 | fail→fail | 18,250 | 7,572 | -59% | 1 | 1 | 0% | 4,280 | 1,251 | -71% | 0 | 0 | — |
case-03 | fail→pass | 8,012 | 11,194 | +40% | 1 | 1 | 0% | 479 | 2,836 | +492% | 0 | 0 | — |
case-09 | fail→pass | 6,994 | 2,200 | -69% | 1 | 1 | 0% | 1,272 | 971 | -24% | 0 | 0 | — |
case-10 | fail→pass | 10,587 | 3,272 | -69% | 1 | 1 | 0% | 1,784 | 931 | -48% | 0 | 0 | — |
case-15 | fail→pass | 17,477 | 1,259 | -93% | 1 | 1 | 0% | 1,681 | 814 | -52% | 0 | 0 | — |
case-04 | fail→fail | 12,290 | 3,082 | -75% | 1 | 1 | 0% | 2,208 | 1,201 | -46% | 0 | 0 | — |
case-05 | fail→pass | 8,123 | 2,993 | -63% | 1 | 1 | 0% | 1,642 | 1,142 | -30% | 0 | 0 | — |
case-06 | fail→pass | 10,201 | 2,598 | -75% | 1 | 1 | 0% | 1,791 | 1,194 | -33% | 0 | 0 | — |
case-07 | fail→pass | 11,610 | 3,947 | -66% | 1 | 1 | 0% | 1,946 | 954 | -51% | 0 | 0 | — |
case-08 | fail→pass | 28,028 | 4,506 | -84% | 1 | 1 | 0% | 2,477 | 1,519 | -39% | 0 | 0 | — |
case-11 | fail→pass | 14,353 | 1,944 | -86% | 1 | 1 | 0% | 2,567 | 967 | -62% | 0 | 0 | — |
case-12 | pass→pass | 5,679 | 3,092 | -46% | 1 | 1 | 0% | 1,170 | 1,159 | -1% | 0 | 0 | — |
case-13 | fail→fail | 3,631 | 2,908 | -20% | 1 | 1 | 0% | 597 | 866 | +45% | 0 | 0 | — |
case-14 | pass→pass | 8,638 | 1,164 | -87% | 1 | 1 | 0% | 1,540 | 766 | -50% | 0 | 0 | — |
case-16 | fail→fail | 7,090 | 1,795 | -75% | 1 | 1 | 0% | 1,136 | 876 | -23% | 0 | 0 | — |
case-17 | fail→fail | 3,970 | 1,393 | -65% | 1 | 1 | 0% | 734 | 776 | +6% | 0 | 0 | — |
case-18 | fail→pass | 11,275 | 1,176 | -90% | 1 | 1 | 0% | 2,018 | 787 | -61% | 0 | 0 | — |
case-19 | pass→pass | 8,670 | 7,920 | -9% | 1 | 1 | 0% | 1,770 | 2,039 | +15% | 0 | 0 | — |
case-20 | pass→pass | 13,704 | 9,312 | -32% | 1 | 1 | 0% | 2,764 | 2,621 | -5% | 0 | 0 | — |
case-21 | pass→pass | 10,268 | 8,872 | -14% | 1 | 1 | 0% | 2,259 | 2,211 | -2% | 0 | 0 | — |
case-22 | pass→pass | 7,655 | 9,770 | +28% | 1 | 1 | 0% | 1,700 | 2,684 | +58% | 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, and 20 counted toward the lift figure. The other 2 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 +50 percentage points is the difference between those two pass rates over the 20 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.