Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Manage gplay's local app registry and app-level details. Use when onboarding packages into gplay, discovering which apps a credential can reach (`apps accessible list`), listing or removing registered apps, pinning a package to the repo, or reading/patching app details (default language, contact info).
.claude/skills/pollyglot-gplay-apps/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | -32% | 0% |
| case-06 | ✗→✓ | ▲ Improved | 0% | 0% |
| case-02 | ✗→✓ | ▲ Improved | -51% | 0% |
| case-03 | ✗→✓ | ▲ Improved | -34% | 0% |
| case-04 | ✗→✓ | ▲ Improved | -53% | 0% |
Two jobs: gplay's local registry of packages, and an app's App details record. Shared conventions are in gplay-cli-usage.
Two different questions, two commands:
apps list: gplay's local registry, the packages you apps add-ed (thePublisher API has no list endpoint).
apps accessible list: what the credential can reach, server-side. The twosets need not coincide: bootstrap from the second, then apps add what you drive.
bashgplay apps accessible list # server-side: apps this credential can reach gplay apps add com.example.app # register one (validates access via the API) gplay apps add com.a com.b com.c # register several: independent, partial success gplay apps list # list packages in the local registry gplay apps view --package com.example.app # default language, title, contact email, icon gplay apps remove com.example.app # drop from the registry (does not touch Play) gplay init # pin a package to ./.gplay for this repo
bashgplay apps details view --package com.example.app gplay apps details set --contact-email support@example.com gplay apps details set --default-language en-US --contact-phone "" # "" clears the field
apps audit: a read-only consistency sweepapps audit ([experimental]) sweeps apps for drift and reports it.
bashgplay apps audit # every app the credential can see gplay apps audit com.example.a com.example.b # only these (skips discovery) gplay apps audit --check lingering-drafts --check empty-release-notes gplay apps audit --skip-check locale-drift --output json
Exit 70 means findings (a gate, not a failure); read the report's ran key before trusting an empty findings.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-20 | pass→pass | 31,320 | 6,702 | -79% | 1 | 1 | 0% | 2,109 | 806 | -62% | 0 | 0 | — |
case-01 | fail→pass | 8,180 | 8,108 | -1% | 1 | 1 | 0% | 1,363 | 933 | -32% | 0 | 0 | — |
case-06 | fail→pass | 13,784 | 11,206 | -19% | 1 | 1 | 0% | 1,207 | 1,201 | -0% | 0 | 0 | — |
case-02 | fail→pass | 19,528 | 5,631 | -71% | 1 | 1 | 0% | 2,314 | 1,128 | -51% | 0 | 0 | — |
case-03 | fail→pass | 86,878 | 6,166 | -93% | 1 | 1 | 0% | 1,530 | 1,008 | -34% | 0 | 0 | — |
case-04 | fail→pass | 22,726 | 9,772 | -57% | 1 | 1 | 0% | 2,552 | 1,188 | -53% | 0 | 0 | — |
case-05 | fail→pass | 14,590 | 2,606 | -82% | 1 | 1 | 0% | 2,176 | 791 | -64% | 0 | 0 | — |
case-07 | fail→pass | 22,367 | 7,449 | -67% | 1 | 1 | 0% | 1,835 | 789 | -57% | 0 | 0 | — |
case-08 | fail→pass | 24,884 | 2,502 | -90% | 1 | 1 | 0% | 4,584 | 798 | -83% | 0 | 0 | — |
case-09 | fail→pass | 44,139 | 7,985 | -82% | 1 | 1 | 0% | 2,266 | 922 | -59% | 0 | 0 | — |
case-10 | pass→pass | 21,385 | 43,929 | +105% | 1 | 1 | 0% | 2,559 | 2,179 | -15% | 0 | 0 | — |
case-11 | fail→pass | 18,037 | 9,842 | -45% | 1 | 1 | 0% | 1,841 | 928 | -50% | 0 | 0 | — |
case-12 | fail→pass | 37,316 | 4,498 | -88% | 1 | 1 | 0% | 2,744 | 1,015 | -63% | 0 | 0 | — |
case-13 | fail→pass | 96,336 | 9,582 | -90% | 1 | 1 | 0% | 2,450 | 810 | -67% | 0 | 0 | — |
case-14 | fail→fail | 12,362 | 11,550 | -7% | 1 | 1 | 0% | 1,686 | 758 | -55% | 0 | 0 | — |
case-15 | fail→pass | 14,887 | 4,174 | -72% | 1 | 1 | 0% | 2,467 | 1,082 | -56% | 0 | 0 | — |
case-16 | fail→pass | 14,688 | 8,222 | -44% | 1 | 1 | 0% | 2,460 | 822 | -67% | 0 | 0 | — |
case-17 | pass→pass | 14,143 | 9,443 | -33% | 1 | 1 | 0% | 1,355 | 953 | -30% | 0 | 0 | — |
case-18 | pass→pass | 20,089 | 3,402 | -83% | 1 | 1 | 0% | 2,169 | 851 | -61% | 0 | 0 | — |
case-19 | pass→pass | 48,027 | 8,239 | -83% | 1 | 1 | 0% | 1,718 | 732 | -57% | 0 | 0 | — |
case-21 | pass→pass | 128,187 | 8,359 | -93% | 1 | 1 | 0% | 1,452 | 871 | -40% | 0 | 0 | — |
case-22 | fail→fail | 17,502 | 11,550 | -34% | 1 | 1 | 0% | 2,100 | 2,279 | +9% | 0 | 0 | — |
case-23 | fail→fail | 8,452 | 12,719 | +50% | 1 | 1 | 0% | 1,367 | 1,727 | +26% | 0 | 0 | — |
case-24 | fail→fail | 10,058 | 13,779 | +37% | 1 | 1 | 0% | 1,088 | 1,172 | +8% | 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 +58 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.
| Model | Method | Date | Lift |
|---|---|---|---|
| gemini-3.6-flash | verified | 9/7/2026 | +68% |
| gemini-3.6-flash | verified | 8/23/2026 | +83% |
| gemini-3.6-flash | verified | 8/4/2026 | +64% |
Other measured skills in the registry, with their headline benchmark lift.