一键导入
test
Run the full LIF Core check suite — ruff, ty, pytest unit, Polylith check, integration, and the advisor-app frontend tests — and report a summary table.
用 Codex 或 Claude 帮你安装 复制这段 Prompt,粘贴到 Codex、Claude 或其他助手里,让它检查 Skill 页面并帮你完成安装。
菜单
Run the full LIF Core check suite — ruff, ty, pytest unit, Polylith check, integration, and the advisor-app frontend tests — and report a summary table.
用 Codex 或 Claude 帮你安装 复制这段 Prompt,粘贴到 Codex、Claude 或其他助手里,让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Triage and respond to review feedback on your open LIF Core PRs — pull each PR's reviews + inline comments, apply fixes (or push back with reasoning), reply inline, and re-request review. Appends commits (never force-pushes) and gates outward-facing replies on your confirmation.
Draft a new LIF Core ADR. Picks the right domain folder, fills that folder's _template.md, assigns the next per-folder number, and commits. Knows the per-domain-folder, per-folder-numbered ADR layout under docs/design/adr/.
Multi-agent staleness sweep of open GitHub issues — fan out cheap agents to check which open issues are already resolved by merged PRs / shipped code, arbitrate the uncertain ones, and surface a close-list with evidence. Closing is gated on user confirmation.
Iteratively design a multi-layer LIF Core feature via a Workflow that runs sequential Opus Plan agents, each refining the prior version through one lens (FP/Polylith → backend correctness → frontend/holistic). Writes v1–v4 plan files to .claude/plans/<feature>-vN.md.
Stand up a new LIF HTTP service end-to-end — Polylith base + project + auth wiring + docker-compose + GitHub Actions + CloudFormation + tests — as an executable checklist with the known pyproject/brick/Docker pitfalls baked in.
Open a convention-compliant LIF Core PR — reads AGENTS.md + CONTRIBUTING.md + the PR template (which do NOT auto-load), verifies branch/commits/pre-commit, fills the template (Closes
| name | test |
| description | Run the full LIF Core check suite — ruff, ty, pytest unit, Polylith check, integration, and the advisor-app frontend tests — and report a summary table. |
| allowed-tools | Bash |
Run the LIF Core quality gate. These are the same checks pre-commit runs (ruff, ty, pytest) plus the integration suite, the Polylith workspace check, and the one frontend with a test runner.
Run from the repo root. The backend checks are independent — run them, then the frontend.
uv run ruff checkuv run ruff format --check (use uv run ruff format to actually fix)uv run ty checkuv run poly check (catches missing brick deps / broken [tool.polylith.bricks] wiring)uv run pytest test/ (mirrors source: test/components/, test/bases/; asyncio_mode = auto)uv run pytest integration_tests/ --skip-unavailablecd frontends/lif_advisor_app && npm test (vitest)Notes:
projects/mongodb/sample_data/{org-key}/. --skip-unavailable skips a test when its backing service is unreachable rather than failing — keep it on for local runs. Full reference: docs/operations/guides/testing.md.mdr-frontend has no test runner (its package.json scripts are dev/build/lint/preview only). Don't try npm test there — for it, fall back to cd frontends/mdr-frontend && npm run build (runs tsc -b) as the type/build gate. (A vitest harness for it is tracked separately.)importlib.reload() to re-run a module under test — it breaks isinstance()/pytest.raises(). Use mock.patch.object(module, "VAR", value).uv run pytest test/components/lif/<component>/.uv run pre-commit run --all-files.Report a summary table:
| Check | Result | Details |
|---|---|---|
| ruff check | pass/fail | lint findings |
| ruff format | pass/fail | files needing format |
| ty check | pass/fail | type errors |
| poly check | pass/fail | brick/dep issues |
| Unit (pytest test/) | pass/fail | N passed / N failed |
| Integration | pass/fail | N passed / N skipped (service unavailable) |
| Advisor-app (vitest) | pass/fail | N tests |
If any check fails, show the relevant error output below the table.