菜单
Defines the workspace, package boundaries, generated artifacts, phase order, compiler gates, development processes, and canonical commands. Read before choosing where to work or which command to run.
默认使用会先检查来源的 Prompt；你也可以切换为直接命令，或下载本地副本。
决定是否安装前，请先阅读 SKILL.md，以及 SkillsMP 当前展示的配套文件。
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。
直接命令不会经过审查 Prompt；运行前请先检查来源。
npx skills add https://github.com/samchon/lint-plugin-evidence --skill project命令会保持在同一行。复制前请横向滚动并检查完整内容。
想先保存到本地？可下载 SkillsMP 当前能够提供的文件。
Defines the review of @evidence and @evidenceExclude for fake citations, for cited requirements the code contradicts, and for exclusions created solely to evade compiler errors, of configuration edited beyond the one permitted activation, and of the compiler and runtime gates each scope must pass. Read only for a review objective; backend.md, frontend.md, and overall.md carry the per-scope configuration and gates.
Defines how an @samchon/lint-plugin-evidence benchmark campaign is set up, launched under frozen inputs, run, supervised, recovered, and reported, from issue creation through pull-request completion. Use whenever operating, supervising, or reporting a benchmark run.
Defines the Evidence Graph tag grammar, truthfulness rules, claim activation, frozen configuration, behavioral proof, citations, exclusions, and the stub marker. Read before Evidence implementation or handling a graph diagnostic; backend.md and frontend.md carry the per-phase claims, placement, examples, and staged unlock.
基于 SOC 职业分类
正在显示 SKILL.md
| name | project |
| description | Defines the workspace, package boundaries, generated artifacts, phase order, compiler gates, development processes, and canonical commands. Read before choosing where to work or which command to run. |
This repository implements every requirement under docs/analysis/ as a NestJS backend, generated Nestia SDK, and React application. The requirement documents are immutable input. A compiling stub, an unreachable endpoint, or an untested behavior is incomplete.
The scaffold already provides package wiring, lint and compiler configuration, SQLite setup, /health, and a frontend shell. Requirement-derived models, contracts, logic, tests, screens, and journeys are yours to implement.
| Path | Owner |
|---|---|
docs/analysis/ | Immutable requirements |
packages/api/src/structures/ | Authored request and response DTOs |
packages/api/src/functional/ | Generated SDK accessors |
packages/backend/prisma/schema/ | Authored Prisma schema |
packages/backend/src/controllers/ | Routes, guards, and delegation |
packages/backend/src/providers/ | Business logic and database access |
packages/backend/test/features/ | Backend end-to-end tests |
packages/frontend/src/ | React application |
packages/frontend/src/lib/<domain>/hooks.ts | The only callers of the generated accessors |
packages/frontend/src/components/<domain>/ | Screens and their components |
packages/frontend/tests/journeys/ | Browser journeys |
Import the API package from its package entry. Do not add a structures subpath export or import.
Do not begin frontend implementation before the backend contract and tests pass. A later frontend finding may reopen the backend only when it identifies a specific requirement, contract, diagnostic, test, or integration defect.
pnpm check:watch from packages/backend with the lifecycle your current objective prescribes. It runs the test/tsconfig.json Program, which compiles the backend tests together with the backend source, and reports lint rules and configured contributors for both. The API package checks its own DTOs.pnpm dev from packages/frontend before frontend authoring. Keep it running through Overall Final. Vite and @ttsc/unplugin report type, lint, and contributor diagnostics on reload.pnpm dev from packages/backend before live frontend integration. Keep it running through Overall Final.If a required development process exits, diagnose its output, fix the owning failure, restart it, and wait for a clean current reload. A stale earlier success is not a gate result.
| Generated artifact | Authored source | Command |
|---|---|---|
packages/backend/src/prisma/** | packages/backend/prisma/schema/** | backend pnpm build:prisma |
docs/ERD.md | packages/backend/prisma/schema/** | backend pnpm build:prisma |
packages/api/src/functional/** | controllers and DTOs | backend pnpm build:sdk |
Never edit generated output. Correct its authored source and regenerate after the complete source change settles.
From packages/backend:
pnpm check:watch
pnpm build:prisma
pnpm schema
pnpm build:sdk
pnpm test
pnpm dev
From packages/frontend:
pnpm dev
pnpm test:e2e
pnpm build:prisma generates the client and ERD. pnpm schema resets the disposable SQLite database. pnpm build:sdk generates the SDK and compiles the API package. pnpm test runs the backend suite. Frontend pnpm test:e2e builds the production bundle and runs browser journeys.
Follow your current objective's instruction for check:watch. Run generators only after their complete authored input settles, and run mutating generators and runtime tests one at a time because they share generated files and the SQLite database.
The workspace-root pnpm build and pnpm test are Overall-phase commands. Do not use a root build to judge an unfinished layer.
Compilation and lint run through ttsc; tests run through ttsx. Do not substitute stock tsc, ts-node, or ESLint. A stock TypeScript build omits the configured contributor rules.
pnpm format rewrites files. It is not a correctness gate.