一键导入
test
Use when adding or evolving tests for a feature/bugfix — applies the sentropic test-pyramid, scoped-test loop, environment isolation, and AI-flaky policy.
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
Use when adding or evolving tests for a feature/bugfix — applies the sentropic test-pyramid, scoped-test loop, environment isolation, and AI-flaky policy.
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
Use when bringing the harness method to a NEW or non-sentropic repo — maps the repo's conventions into a harness profile so the same scope/branch/verify kernel applies anywhere.
Use before any creative/design work (a feature, a package, an evolution) — explores intent and design via the spec-ladder with multi-peer adversarial review BEFORE implementation.
Use on any bug, test failure, or unexpected behaviour BEFORE proposing a fix — runs the generic evidence-first root-cause loop.
Use when you have a spec/decisions and need an executable plan — writes a lot-based BRANCH.md from the template and completes with track for realization state.
Use when completing a feature, a design step, or before merging — runs a ≥2-peer consensus review and reconciles the findings, instead of a single rubber-stamp pass.
Load FIRST in any sentropic repo — establishes the harness method (branch/scope/test/debug/review/brainstorm/plan) and supersedes superpowers for those acts.
| name | test |
| description | Use when adding or evolving tests for a feature/bugfix — applies the sentropic test-pyramid, scoped-test loop, environment isolation, and AI-flaky policy. |
Native sentropic testing discipline (NOT "tdd" ritual — the act is test). Open it with
harness test [<scope>] [--category unit|integration|e2e] [--watch] (records a WorkEvent), then follow
the loop.
70% unit · 20% integration · 10% e2e. Test behaviour, not implementation — query by user-visible elements (text/role/label) over test ids. Names: "should when ".
Run ONLY the test under evolution, on a dedicated test ENV — never dev:
make test-api-<suite> SCOPE=tests/<file>.spec.ts ENV=test-<slug>make test-ui SCOPE=tests/<file>.spec.ts ENV=test-<slug>make test-e2e E2E_SPEC=tests/<file>.spec.ts API_PORT=… UI_PORT=… MAILDEV_UI_PORT=… ENV=e2e-<slug>Full gates only at end of lot/branch: make test-api, make test-ui, make clean test-e2e.
ENV=dev — afterEach hooks purge real data.ENV=test-<slug> / ENV=e2e-<slug>; ENV is the LAST
make argument.Only accept non-systematic provider/network/model nondeterminism, and only with ≥1 success on the same
commit + command. Record the exact command + failing file + failure signature in BRANCH.md and get
explicit user sign-off before merge. Everything else is blocking.
CI was green on main → any branch failure IS a branch problem; never claim "pre-existing". Decide: evolution (adapt the test) or regression (fix the code). Local-green / CI-red → trust CI.