메뉴
Diagnose common Rockets SDK startup/build/security wiring issues with executable checks and fix-ready output. Use when an app fails to boot, build, or enforce ACL as expected.
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
SOC 직업 분류 기준
This skill should be used when adding ACL to an existing module, modifying role-based access rules, debugging 403 errors, configuring ownership checks, or writing Access Query Services manually. For new modules, use rockets-crud-generator with ACL config instead — it generates all ACL wiring automatically.
This skill should be used when generating an entire Rockets SDK project from a spec, bootstrapping multiple entities at once, or when the user has a plan.json file. Orchestrates wave-based generation with dependency ordering and validation gates. Works sequentially by default; supports Agent Teams for parallel execution. Requires rockets-crud-generator.
This skill should be used when implementing business logic beyond standard CRUD — state machines, approval workflows, event-driven automation, notifications, file uploads, external API integration, or cross-entity orchestration. Use after CRUD modules are generated. Also covers the foundational rule that application services must inject model services, never repositories.
This skill should be used when the user asks to generate a CRUD module, create a new entity, scaffold a domain object, add a new resource with endpoints, or create a junction table. Generates complete Rockets SDK modules including TypeORM entities, NestJS modules, controllers, services, DTOs, interfaces, and ACL wiring using generate.js, integrate.js, and validate.js scripts.
This skill should be used when starting a new Rockets SDK project, setting up a fresh API, or initializing from the rockets-starter template. Automates clone, env setup, dependency install, and build/test verification.
| name | rockets-runtime-diagnostics |
| description | Diagnose common Rockets SDK startup/build/security wiring issues with executable checks and fix-ready output. Use when an app fails to boot, build, or enforce ACL as expected. |
Use this skill to diagnose real runtime/build problems in Rockets projects, not just read guides.
build and test commands.# Diagnose project in current directory
node skills/rockets-runtime-diagnostics/scripts/diagnose.js .
# Diagnose another project
node skills/rockets-runtime-diagnostics/scripts/diagnose.js /path/to/project
# Include build and tests
node skills/rockets-runtime-diagnostics/scripts/diagnose.js . --run-build --run-tests
# Machine-readable output
node skills/rockets-runtime-diagnostics/scripts/diagnose.js . --json
ROCKETS_AUTH_IMPORT_ORDER: RocketsAuthModule must appear before RocketsModule/RocketsServerModule.ROCKETS_MODULE_NOT_REGISTERED: package installed but module not registered in app.module.ts.DYNAMIC_USER_METADATA_TOKEN_MISSING: userMetadata configured without TypeOrmExtModule.forFeature({ userMetadata: ... }).ACL_RESOURCE_LITERAL_NOT_IN_ENUM: literal resource in decorators missing in AppResource enum.ACCESS_QUERY_SERVICE_MISSING: module has CRUD controller but no *-access-query.service.ts.yarn build and then targeted tests.Tests live CRUD endpoints against a running NestJS app:
# Basic usage
node skills/rockets-runtime-diagnostics/scripts/smoke-test-endpoints.js --project ./apps/api
# Custom port + longer timeout
node skills/rockets-runtime-diagnostics/scripts/smoke-test-endpoints.js --project ./apps/api --port 4000 --timeout 60000
# Without auth (public endpoints only)
node skills/rockets-runtime-diagnostics/scripts/smoke-test-endpoints.js --project ./apps/api --no-auth
# Keep server running after tests
node skills/rockets-runtime-diagnostics/scripts/smoke-test-endpoints.js --project ./apps/api --keep
What it does:
src/modules/*/\*.crud.controller.tsyarn start or npm start)Output: { passed, summary: { total, passed, failed, modules }, auth, modules }
node skills/rockets-runtime-diagnostics/scripts/smoke-test.js
This validates the diagnostic engine against fixture projects with known errors.