ワンクリックで
team-run-tests
Batch test and fix — run comprehensive tests, identify all issues, then fix systematically
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
メニュー
Batch test and fix — run comprehensive tests, identify all issues, then fix systematically
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
SOC 職業分類に基づく
| name | team-run-tests |
| description | Batch test and fix — run comprehensive tests, identify all issues, then fix systematically |
| argument-hint | [test pattern or scope] |
| disable-model-invocation | true |
| allowed-tools | Bash(git *), TaskCreate, TaskUpdate, TaskList, TaskGet |
Run all tests, batch-fix failures, and verify. Loops until green.
Test scope: $ARGUMENTS
Enable with CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS=1
Create a team and spawn teammates:
test-runCreate the following task list with dependencies:
| # | Task | Owner | Blocked By |
|---|---|---|---|
| 1 | Create branch test/<timestamp> from main | shipper | — |
| 2 | Run full test suite, compile failure report | shipper | 1 |
| 3 | Fix application test failures (dynamic sub-tasks per failure) | full-stack-developer | 2 |
| 4 | Fix data layer test failures (dynamic sub-tasks per failure) | database-admin | 2 |
| 5 | Commit all fixes | shipper | 3, 4 |
| 6 | Re-run previously failed tests | shipper | 5 |
| 7 | Run full validation suite | shipper | 6 |
| 8 | Review all fixes for quality | reviewer | 7 |
| 9 | Update documentation if fixes affect docs | documentor | 8 |
| 10 | Create PR to main with test report | shipper | 8, 9 |
Parallelism: Tasks 3 and 4 can run simultaneously.
Loop: If task 6 finds remaining failures, create new fix tasks and loop back to task 6.
Dynamic tasks: After task 2, create specific sub-tasks for each failure group and assign to the appropriate developer.
Execute sequentially using the Task tool:
Task(shipper, "Create branch test/<timestamp> from main")Task(shipper, "Run full test suite, report all failures")Task(full-stack-developer, "Fix: <failure details>") — for app failuresTask(database-admin, "Fix: <failure details>") — for data failuresTask(shipper, "Commit fixes: fix: <test-name> - <description>")Task(shipper, "Re-run previously failed tests")Task(shipper, "Run full test suite for final validation")Task(reviewer, "Review all test fixes")Task(documentor, "Update docs if test fixes changed documented behavior")Task(shipper, "Create PR to main with test report") Phase 1: Discovery Phase 2: Fix Phase 3: Verify
┌─────────┐ ┌─────────┐ ┌──────────────┐ ┌─────────┐ ┌──────────┐
│ Shipper │────►│ Shipper │────►│Full Stack Dev│────►│ Shipper │────►│ Shipper │
│ Branch │ │Run Tests│ │ + DB Admin │ │ Commit │ │ Re-test │
└─────────┘ └────┬────┘ └──────────────┘ └─────────┘ └────┬─────┘
│ ↑ │
▼ │ (loop if ▼
[Failure Report]────────┘ still failing) [Full Validation]
│
▼
┌──────────┐ ┌────────────┐
│ Reviewer │────►│ Documentor │──►[PR/Merge]
└──────────┘ │ Update Docs│
└────────────┘
Build and deploy features — branch, implement, commit, review, test, deploy, PR/merge
shadcn/ui component patterns including Radix UI primitives, CVA variants, form components with React Hook Form, and Tailwind CSS styling. Use when creating UI components, building forms, implementing design systems, or working with shadcn/ui.
Add critical test coverage — tests ONLY for functionality that could cause production disasters
List all available tools with their signatures and purposes
Analyze Claude Code audit logs — session summaries, metrics, timelines, anomaly detection, and claim verification
Technical debt and refactoring — analyze, refactor, and validate code improvements