Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Inspect and manage Google Play release tracks and their testers. Use when checking what's on a track, creating a closed-testing track before an upload, auditing country availability (read-only), or setting the Google Groups authorized to test a closed track.
.claude/skills/pollyglot-gplay-tracks/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | -68% | 0% |
| case-02 | ✗→✓ | ▲ Improved | -64% | 0% |
| case-03 | ✗→✓ | ▲ Improved | -74% | 0% |
| case-04 | ✗→✓ | ▲ Improved | -70% | 0% |
| case-06 | ✗→✓ | ▲ Improved | -50% | 0% |
Shared conventions are in gplay-cli-usage; shipping builds onto tracks is gplay-release-flow.
bashgplay tracks list --package com.example.app # every track on the app gplay tracks view --track production # one track's full state
bashgplay tracks create qa-team
Creates a closed-testing track; open/internal tracks have no API path (they always exist). An upload or promote to a missing custom track fails with exit 30: create it here first (gplay-release-flow, trackhint).
bashgplay tracks availability view --track production
Read-only at the API level (ADR-0012); changing availability is a Play Console job.
bashgplay testers list --track qa-team gplay testers set --track qa-team --group qa@googlegroups.com,beta@googlegroups.com gplay testers set --track qa-team --clear # close the closed test
testers set replaces the whole audience (no add/remove) and takes Google Groups only.
bashgplay tracks create qa-team gplay testers set --track qa-team --group qa@googlegroups.com gplay releases upload ./app.aab --track qa-team # see gplay-release-flow
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 17,135 | 9,572 | -44% | 1 | 1 | 0% | 2,856 | 902 | -68% | 0 | 0 | — |
case-02 | fail→pass | 18,714 | 9,000 | -52% | 1 | 1 | 0% | 2,544 | 912 | -64% | 0 | 0 | — |
case-03 | fail→pass | 47,416 | 3,077 | -94% | 1 | 1 | 0% | 2,894 | 746 | -74% | 0 | 0 | — |
case-04 | fail→pass | 19,908 | 8,271 | -58% | 1 | 1 | 0% | 2,387 | 727 | -70% | 0 | 0 | — |
case-05 | pass→pass | 19,836 | 10,065 | -49% | 1 | 1 | 0% | 2,410 | 980 | -59% | 0 | 0 | — |
case-06 | fail→pass | 18,749 | 5,504 | -71% | 1 | 1 | 0% | 2,059 | 1,025 | -50% | 0 | 0 | — |
case-07 | pass→pass | 20,781 | 23,265 | +12% | 1 | 1 | 0% | 2,202 | 929 | -58% | 0 | 0 | — |
case-08 | fail→pass | 13,761 | 9,474 | -31% | 1 | 1 | 0% | 2,072 | 993 | -52% | 0 | 0 | — |
case-09 | fail→pass | 14,481 | 8,337 | -42% | 1 | 1 | 0% | 2,287 | 785 | -66% | 0 | 0 | — |
case-10 | fail→pass | 18,573 | 7,300 | -61% | 1 | 1 | 0% | 2,307 | 581 | -75% | 0 | 0 | — |
case-11 | fail→pass | 25,354 | 2,812 | -89% | 1 | 1 | 0% | 2,219 | 658 | -70% | 0 | 0 | — |
case-12 | fail→pass | 46,633 | 3,557 | -92% | 1 | 1 | 0% | 2,998 | 814 | -73% | 0 | 0 | — |
case-13 | fail→pass | 16,581 | 6,933 | -58% | 1 | 1 | 0% | 2,019 | 714 | -65% | 0 | 0 | — |
case-14 | fail→pass | 15,262 | 7,504 | -51% | 1 | 1 | 0% | 2,375 | 602 | -75% | 0 | 0 | — |
case-15 | fail→pass | 21,733 | 11,063 | -49% | 1 | 1 | 0% | 2,712 | 582 | -79% | 0 | 0 | — |
case-16 | fail→pass | 22,927 | 7,531 | -67% | 1 | 1 | 0% | 2,543 | 628 | -75% | 0 | 0 | — |
case-17 | fail→pass | 53,644 | 8,188 | -85% | 1 | 1 | 0% | 2,694 | 820 | -70% | 0 | 0 | — |
case-18 | fail→pass | 36,532 | 5,889 | -84% | 1 | 1 | 0% | 5,595 | 812 | -85% | 0 | 0 | — |
case-19 | pass→pass | 17,006 | 4,471 | -74% | 1 | 1 | 0% | 1,903 | 953 | -50% | 0 | 0 | — |
case-20 | fail→pass | 21,029 | 8,266 | -61% | 1 | 1 | 0% | 2,506 | 699 | -72% | 0 | 0 | — |
case-21 | pass→pass | 26,943 | 9,469 | -65% | 1 | 1 | 0% | 1,183 | 856 | -28% | 0 | 0 | — |
case-22 | fail→pass | 26,406 | 3,397 | -87% | 1 | 1 | 0% | 2,792 | 694 | -75% | 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 +82 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.
| Model | Method | Date | Lift |
|---|---|---|---|
| gemini-3.6-flash | verified | 8/23/2026 | +55% |
| gemini-3.6-flash | verified | 8/4/2026 | +59% |
Other measured skills in the registry, with their headline benchmark lift.