Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Configure a game's Play Games Services achievements and leaderboards with gplay `games`, keyed by the numeric `--application-id`; writes touch the draft only (publishing to players is Console-only). Use when creating or editing achievements or leaderboards, scripting their config in CI, or round-tripping a configuration as JSON.
.claude/skills/pollyglot-gplay-games/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | -34% | 0% |
| case-02 | ✗→✓ | ▲ Improved | -30% | 0% |
| case-03 | ✗→✓ | ▲ Improved | -28% | 0% |
| case-04 | ✗→✓ | ▲ Improved | 1% | 0% |
| case-05 | ✗→✓ | ▲ Improved | -17% | 0% |
gplay games configures a game's Play Games Services achievement and leaderboard configurations. Shared conventions are in gplay-cli-usage. The whole namespace is [experimental].
--application-idPlay Games resources are keyed by the numeric Play Games application ID, not the package: --application-id on list/create, the resource id alone on view/update/delete. The .gplay/config.json pin plays no part here.
bashgplay games achievements list --application-id 1234567890 gplay games leaderboards list --application-id 1234567890
Every write edits the draft. There is no publish method: create/update stage the change, then a human publishes it in the Play Console. Say so when handing off.
bashgplay games achievements view <achievementId> --output json # read one gplay games achievements create --application-id 123 \ --name "First Blood" --description "Win your first match" \ --type STANDARD --initial-state REVEALED --point-value 10
Field flags, or --from-json for a full body: the round-trip of view --output json, and the way to set several locales at once.
bashgplay games leaderboards view <leaderboardId> --output json # read one gplay games leaderboards create --application-id 123 \ --name "High Scores" --score-order LARGER_IS_BETTER \ --score-min 0 --score-max 1000000 gplay games leaderboards update <leaderboardId> --name "Top Scores"
update replaces: fetch, edit, resendupdate <id> is a full PUT replace: fetch, edit, resend with --from-json. The field flags alone are fine for a one-field change.
bashgplay games achievements view <id> --output json > ach.json # edit ach.json … gplay games achievements update <id> --from-json ach.json
create/update are routine draft writes (--dry-run previews the request body). delete is irreversible and sits on the --confirm tier:
bashgplay games achievements delete <achievementId> --confirm gplay games leaderboards delete <leaderboardId> --confirm
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 13,790 | 9,126 | -34% | 1 | 1 | 0% | 1,812 | 1,202 | -34% | 0 | 0 | — |
case-02 | fail→pass | 15,502 | 6,193 | -60% | 1 | 1 | 0% | 1,348 | 940 | -30% | 0 | 0 | — |
case-03 | fail→pass | 17,461 | 7,207 | -59% | 1 | 1 | 0% | 2,204 | 1,586 | -28% | 0 | 0 | — |
case-04 | fail→pass | 27,997 | 6,315 | -77% | 1 | 1 | 0% | 1,358 | 1,371 | +1% | 0 | 0 | — |
case-05 | fail→pass | 10,406 | 5,430 | -48% | 1 | 1 | 0% | 1,576 | 1,309 | -17% | 0 | 0 | — |
case-06 | fail→pass | 17,611 | 12,119 | -31% | 1 | 1 | 0% | 1,981 | 1,733 | -13% | 0 | 0 | — |
case-07 | fail→pass | 13,563 | 9,501 | -30% | 1 | 1 | 0% | 1,346 | 894 | -34% | 0 | 0 | — |
case-08 | fail→pass | 9,497 | 7,348 | -23% | 1 | 1 | 0% | 1,541 | 855 | -45% | 0 | 0 | — |
case-09 | fail→pass | 7,200 | 8,863 | +23% | 1 | 1 | 0% | 1,100 | 884 | -20% | 0 | 0 | — |
case-10 | fail→pass | 13,907 | 2,869 | -79% | 1 | 1 | 0% | 1,463 | 876 | -40% | 0 | 0 | — |
case-11 | fail→pass | 14,402 | 5,514 | -62% | 1 | 1 | 0% | 1,742 | 1,122 | -36% | 0 | 0 | — |
case-12 | fail→pass | 9,020 | 13,718 | +52% | 1 | 1 | 0% | 1,317 | 1,004 | -24% | 0 | 0 | — |
case-13 | fail→pass | 9,100 | 8,117 | -11% | 1 | 1 | 0% | 1,233 | 961 | -22% | 0 | 0 | — |
case-14 | fail→pass | 26,236 | 20,612 | -21% | 1 | 1 | 0% | 1,828 | 1,391 | -24% | 0 | 0 | — |
case-15 | fail→pass | 19,045 | 10,712 | -44% | 1 | 1 | 0% | 2,152 | 1,399 | -35% | 0 | 0 | — |
case-16 | fail→pass | 14,731 | 12,804 | -13% | 1 | 1 | 0% | 2,378 | 1,816 | -24% | 0 | 0 | — |
case-17 | fail→pass | 15,605 | 9,028 | -42% | 1 | 1 | 0% | 1,557 | 1,154 | -26% | 0 | 0 | — |
case-18 | pass→pass | 25,185 | 10,425 | -59% | 1 | 1 | 0% | 2,255 | 1,554 | -31% | 0 | 0 | — |
case-19 | pass→pass | 20,354 | 9,252 | -55% | 1 | 1 | 0% | 2,284 | 1,168 | -49% | 0 | 0 | — |
case-20 | fail→pass | 13,342 | 25,868 | +94% | 1 | 1 | 0% | 1,504 | 1,226 | -18% | 0 | 0 | — |
case-21 | pass→pass | 44,298 | 8,797 | -80% | 1 | 1 | 0% | 1,503 | 980 | -35% | 0 | 0 | — |
case-22 | fail→pass | 20,808 | 7,911 | -62% | 1 | 1 | 0% | 2,161 | 973 | -55% | 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 +86 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 | 9/7/2026 | +61% |
| gemini-3.6-flash | verified | 8/23/2026 | +82% |
| gemini-3.6-flash | verified | 8/4/2026 | +59% |
Other measured skills in the registry, with their headline benchmark lift.