Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Google Cloud Platform (GCP) の Application Default Credentials 認証を実行するスキル。 「GCP認証して」「Google Cloud認証」「gcloud login」等のリクエストで発動。 BigQuery や Cloud Storage 等の GCP サービス利用前の認証手順をガイド。
.claude/skills/minicoohei-gcp-auth/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-18 | ✗→✓ | ▲ Improved | -25% | 0% |
| case-01 | ✓→✓ | ＝ Same ✓ | 2% | 0% |
| case-02 | ✓→✓ | ＝ Same ✓ | -31% | 0% |
| case-03 | ✓→✓ | ＝ Same ✓ | 29% | 0% |
| case-04 | ✓→✓ | ＝ Same ✓ | -20% | 0% |
「GCP認証」「Google Cloud認証」「gcloud login」「サービスアカウント」
Google Cloud の Application Default Credentials (ADC) 認証を実行するスキルです。
bashgcloud auth application-default login
注意: このコマンドはブラウザでの認証が必要なため、ターミナルで直接実行してください。
| エラー | 対処法 | |--------|--------| | Reauthentication is needed | 認証期限切れ。再度 gcloud auth application-default login を実行 | | GOOGLE_APPLICATION_CREDENTIALS 警告 | .env から該当行を削除するか、ADC を使用 | | Project not set | gcloud config set project PROJECT_ID でプロジェクト設定 |
bash# トークンが表示されれば認証済み gcloud auth application-default print-access-token # 現在のプロジェクト確認 gcloud config get-value project
gcloud) がインストール済み| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | pass→pass | 5,897 | 4,039 | -32% | 1 | 1 | 0% | 990 | 1,014 | +2% | 0 | 0 | — |
case-02 | pass→pass | 9,242 | 4,706 | -49% | 1 | 1 | 0% | 1,601 | 1,098 | -31% | 0 | 0 | — |
case-03 | pass→pass | 3,684 | 3,279 | -11% | 1 | 1 | 0% | 576 | 744 | +29% | 0 | 0 | — |
case-04 | pass→pass | 5,780 | 2,659 | -54% | 1 | 1 | 0% | 944 | 753 | -20% | 0 | 0 | — |
case-05 | pass→pass | 9,456 | 3,656 | -61% | 1 | 1 | 0% | 1,566 | 888 | -43% | 0 | 0 | — |
case-06 | pass→pass | 9,991 | 5,460 | -45% | 1 | 1 | 0% | 1,699 | 1,240 | -27% | 0 | 0 | — |
case-07 | pass→pass | 5,286 | 2,339 | -56% | 1 | 1 | 0% | 978 | 700 | -28% | 0 | 0 | — |
case-08 | pass→pass | 10,551 | 2,663 | -75% | 1 | 1 | 0% | 1,491 | 687 | -54% | 0 | 0 | — |
case-09 | pass→pass | 5,398 | 2,426 | -55% | 1 | 1 | 0% | 881 | 656 | -26% | 0 | 0 | — |
case-10 | pass→pass | 4,273 | 3,001 | -30% | 1 | 1 | 0% | 707 | 622 | -12% | 0 | 0 | — |
case-11 | pass→pass | 2,610 | 12,511 | +379% | 1 | 1 | 0% | 288 | 606 | +110% | 0 | 0 | — |
case-12 | pass→pass | 4,760 | 3,448 | -28% | 1 | 1 | 0% | 842 | 911 | +8% | 0 | 0 | — |
case-13 | pass→pass | 9,739 | 7,769 | -20% | 1 | 1 | 0% | 1,641 | 1,728 | +5% | 0 | 0 | — |
case-14 | pass→pass | 10,037 | 6,191 | -38% | 1 | 1 | 0% | 1,789 | 1,438 | -20% | 0 | 0 | — |
case-15 | pass→pass | 5,878 | 1,789 | -70% | 1 | 1 | 0% | 924 | 598 | -35% | 0 | 0 | — |
case-16 | pass→pass | 1,830 | 1,799 | -2% | 1 | 1 | 0% | 264 | 674 | +155% | 0 | 0 | — |
case-17 | pass→pass | 6,837 | 2,766 | -60% | 1 | 1 | 0% | 1,039 | 776 | -25% | 0 | 0 | — |
case-18 | fail→pass | 9,710 | 5,260 | -46% | 1 | 1 | 0% | 1,689 | 1,269 | -25% | 0 | 0 | — |
case-19 | pass→pass | 11,826 | 3,792 | -68% | 1 | 1 | 0% | 1,875 | 921 | -51% | 0 | 0 | — |
case-20 | pass→pass | 7,298 | 5,799 | -21% | 1 | 1 | 0% | 1,326 | 1,416 | +7% | 0 | 0 | — |
case-21 | pass→pass | 14,064 | 14,662 | +4% | 1 | 1 | 0% | 2,609 | 3,430 | +31% | 0 | 0 | — |
case-22 | pass→pass | 7,051 | 4,096 | -42% | 1 | 1 | 0% | 1,206 | 1,180 | -2% | 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 +5 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.
Other measured skills in the registry, with their headline benchmark lift.