Add AGENTS.md and SKILL.md to support AI tools (#393)

Signed-off-by: dongjiang <dongjiang1989@126.com>
This commit is contained in:
dongjiang 2026-05-12 15:35:17 +08:00 committed by GitHub
parent 15e55e8055
commit 41b1d03cfc
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
11 changed files with 1665 additions and 1 deletions

View file

@ -0,0 +1,140 @@
---
name: api-and-namespace-design
description: API design conventions, namespace coordinate system, RBAC roles, ClawHub compatibility layer, OpenAPI contract sync rules, and CSRF/session handling.
license: Apache-2.0
---
# API and Namespace Design Skill
## Trigger
Use this skill when:
- Adding or modifying REST API endpoints
- Changing namespace, skill, or user coordinate logic
- Working on ClawHub CLI compatibility layer
- Modifying OpenAPI specifications or generated types
- Adding new admin or governance endpoints
## Namespace Coordinate System
SkillHub uses a two-axis coordinate model:
```
@{namespace_slug}/{skill_slug}
```
- `@global/my-skill` — Global namespace skill
- `@my-team/my-skill` — Team namespace skill (namespace slug is any valid slug)
- `@department-ops/my-skill` — Department namespace skill
### Namespace Model
Namespaces (`domain/namespace/`):
- **Slug**: unique identifier, validated by `SlugValidator`
- **Status**: `ACTIVE`, `FROZEN`, `ARCHIVED`
- **Roles**: `OWNER`, `ADMIN`, `MEMBER`
- Frozen or archived namespaces cannot publish skills
### RBAC Roles
**Namespace-level** (`domain/namespace/NamespaceRole`):
- `OWNER` — Full control over namespace and all skills
- `ADMIN` — Can manage members, archive skills, publish
- `MEMBER` — Can publish skills to the namespace
**Platform-level**:
- `SUPER_ADMIN` — Bypasses all permission checks, can publish directly without review
## ClawHub Compatibility Layer
ClawHub CLI uses a single-slug model (no `/` allowed in slugs). Mapping:
| SkillHub Coordinate | Canonical Slug | Notes |
|---------------------|----------------|-------|
| `@global/my-skill` | `my-skill` | Global namespace omits prefix |
| `@team-name/my-skill` | `team-name--my-skill` | Double-dash separator |
**Conflict resolution**: `--` split takes priority. `@global/team-name--my-skill` would conflict
with `@team-name/my-skill`, resolved to the team namespace skill. Global skill slugs must NOT
contain `--`.
## API Design
### Controllers
- Controllers in `skillhub-app` (`com.iflytek.skillhub.controller/`) are **transport only**
- Responsibilities: extract auth context, bind request params, wrap responses
- Complex business logic belongs in domain services (`skillhub-domain`) or app services
- Use Springdoc OpenAPI annotations (`@Operation`, `@ApiResponse`) for API documentation
- User identity is always **String** in API inputs and outputs
### Request/Response Patterns
- DTOs in `com.iflytek.skillhub.dto/`
- `ReviewTaskRequest` / `ReviewTaskResponse` for review workflow
- Response wrapping handled at controller layer
- Validation errors use `DomainBadRequestException` with i18n message keys
### Session and CSRF
- Session-based auth with cookie storage
- CSRF protection via `XSRF-TOKEN` cookie and `X-XSRF-TOKEN` header
- Smoke tests validate the full register → login → CSRF → action → logout flow
- Mock auth uses `X-Mock-User-Id` header in local dev
### Well-known Discovery
`/.well-known/clawhub.json` returns `{ "apiBase": "/api/v1" }` for ClawHub CLI auto-discovery.
## OpenAPI Contract Sync
When backend API contracts change:
```bash
make generate-api
```
This runs `openapi-typescript http://localhost:8080/v3/api-docs -o src/api/generated/schema.d.ts`.
Commit the updated `web/src/api/generated/schema.d.ts` with the PR.
To verify no drift:
```bash
./scripts/check-openapi-generated.sh
```
This starts local dependencies, boots the backend, regenerates the schema, and fails if the
checked-in SDK is stale.
## Versioning and Tags
- Semantic versioning for skill versions (`major.minor.patch`)
- `latest` tag is system-reserved, read-only, auto-follows `Skill.latestVersionId`
- Custom tags (`stable`, `beta`) are manually maintained
- `latest` cannot be moved manually
- Auto-generated versions use `yyyyMMdd.HHmmss` format when no version is specified in SKILL.md
## Key API Endpoints
| Method | Path | Purpose |
|--------|------|---------|
| `GET` | `/api/v1/auth/me` | Current user info (401 if unauthenticated) |
| `POST` | `/api/v1/auth/local/login` | Local account login |
| `POST` | `/api/v1/auth/local/register` | Local account registration |
| `POST` | `/api/v1/auth/logout` | Logout (302/200/204) |
| `POST` | `/api/v1/auth/local/change-password` | Password change |
| `GET` | `/api/v1/namespaces` | List namespaces |
| `GET` | `/api/v1/labels` | List visible labels (public) |
| `POST` | `/api/v1/admin/labels` | Create label definition (admin) |
| `DELETE` | `/api/v1/admin/labels/{slug}` | Delete label definition (admin) |
| `GET` | `/actuator/health` | Health check |
| `GET` | `/actuator/prometheus` | Prometheus metrics |
## Common Pitfalls
- Forgetting CSRF token on POST/PUT/DELETE requests (needs `X-XSRF-TOKEN` header)
- Using numeric user IDs in API — all user identities are **String**
- Not regenerating OpenAPI types after adding/changing endpoints
- Putting business logic in controllers instead of domain/app services
- Assuming namespace slugs follow a specific prefix pattern — they are arbitrary valid slugs

View file

@ -0,0 +1,108 @@
---
name: backend-module-structure
description: Rules for the SkillHub backend Maven multi-module clean architecture. Ensures agents place new code in the correct module and respect dependency direction.
license: Apache-2.0
---
# Backend Module Structure Skill
## Trigger
Use this skill when:
- Adding or modifying Java backend code
- Creating new services, controllers, repositories, or entities
- Refactoring backend code across files
- Reviewing backend code placement
## Rules
### Dependency Direction
The design-doc dependency direction:
```
app → domain, auth, search, storage, infra, notification
infra → domain # implements domain repository interfaces
auth → domain
search → domain
notification → domain
storage → (independent) # pure SPI
```
**Design intent**: `skillhub-domain` should be the innermost layer, defining entities,
repository interfaces, and domain services without depending on infra, auth, search, or storage.
**Code reality**: `skillhub-domain` declares a Maven dependency on `skillhub-storage` (via
`pom.xml`), and several domain services (`SkillHardDeleteService`, `SkillDownloadService`,
`SkillPublishService`, `SkillGovernanceService`, `SkillQueryService`,
`SkillStorageDeletionCompensationService`) import `com.iflytek.skillhub.storage.ObjectStorageService`.
This is an existing deviation from the ideal clean architecture. New code should avoid adding
further cross-module dependencies from domain.
### Where to Place Code
| Code Type | Module | Java Package |
|-----------|--------|-------------|
| Entity / Value Object | skillhub-domain | `com.iflytek.skillhub.domain.{submodule}/` |
| Repository Interface | skillhub-domain | `com.iflytek.skillhub.domain.{submodule}/` |
| Domain Service | skillhub-domain | `com.iflytek.skillhub.domain.{submodule}/service/` |
| Domain Event | skillhub-domain | `com.iflytek.skillhub.domain/event/` |
| Domain Exception | skillhub-domain | `com.iflytek.skillhub.domain/shared/exception/` |
| JPA Repository Impl | skillhub-infra | `com.iflytek.skillhub.infra.repository/` |
| Controller | skillhub-app | `com.iflytek.skillhub.controller/` |
| App Service | skillhub-app | `com.iflytek.skillhub.service/` |
| Query Repository | skillhub-app | `com.iflytek.skillhub.repository/` |
| DTO / Response | skillhub-app | `com.iflytek.skillhub.dto/` |
| OAuth2 / Auth Config | skillhub-auth | `com.iflytek.skillhub.auth/` |
| Search SPI / Impl | skillhub-search | `com.iflytek.skillhub.search/` |
| Storage SPI / Impl | skillhub-storage | `com.iflytek.skillhub.storage/` |
| Notification Service | skillhub-notification | `com.iflytek.skillhub.notification/` |
### Maven Modules
The parent POM (`server/pom.xml`) defines 7 modules with `spring-boot-starter-parent:3.2.3`:
```
skillhub-app | skillhub-domain | skillhub-auth | skillhub-search
skillhub-storage | skillhub-infra | skillhub-notification
```
### Repository vs Query Repository
- **Domain Repository** (`skillhub-domain`): Aggregate reads, state transitions, rule evaluation.
Returns domain objects. Defined as interfaces, implemented in `skillhub-infra` via Spring Data JPA.
- **Query Repository** (`com.iflytek.skillhub.repository`): Read-model assembly, joins multiple
sources, presentation projection. Returns DTOs. Implemented directly in `skillhub-app`.
Current query repositories:
- `GovernanceQueryRepository` / `JpaGovernanceQueryRepository`
- `MySkillQueryRepository` / `JpaMySkillQueryRepository`
- `ProfileReviewQueryRepository` / `JpaProfileReviewQueryRepository`
- `AdminSkillReportQueryRepository` / `JpaAdminSkillReportQueryRepository`
When a new read use case arrives:
1. If it's for state transition or domain rule → domain repository port
2. If it's for page/list/detail response assembly with joins → app query repository
3. If it's a thin single-aggregate read → direct domain repository call from app service
4. If direct SQL/EntityManager is needed → add class-level comment explaining why
### Building Backend Tests
Never run `./mvnw -pl skillhub-app clean test` directly under `server/`. Use:
```bash
make test-backend-app # skillhub-app + dependencies (includes -am)
make test-backend # all backend modules
```
Running clean test on skillhub-app alone can fall back to stale artifacts from the local Maven
repository, surfacing misleading `cannot find symbol` and signature-mismatch errors.
### User Identity Type
User identity is **always String** throughout the codebase. This covers:
- Authentication, API params, permissions, audit
- Resource owner, creator, reviewer, actor, submittedBy
- All user-associated fields
The `UserAccount` entity uses `@Column(length = 128)` for its ID. The platform needs to support
external SSO/OIDC/SCIM identity sources whose UIDs are typically stable strings.

View file

@ -0,0 +1,135 @@
---
name: code-conventions
description: Code style, logging, and testing conventions for SkillHub backend (Java) and frontend (TypeScript). Use when writing or reviewing code.
license: Apache-2.0
---
# Code Conventions Skill
## Java / Backend Conventions
### User Identity Type
User identity is **always `String`** throughout the codebase. This covers:
- Authentication and authorization
- API parameters and responses
- Permission checks
- Audit logs
- Resource owner, creator, reviewer, actor, submittedBy fields
Never introduce `int`, `long`, or `bigint` as user identifiers. The platform needs to support
external SSO/OIDC/SCIM identity sources whose UIDs are typically stable strings.
### Exception Handling
- Use `LocalizedDomainException` for user-facing error messages (supports i18n)
- Use `DomainBadRequestException` for invalid client input
- Use `DomainNotFoundException` for missing resources
- Use `DomainForbiddenException` for authorization failures
- Exception classes live in `skillhub-domain/shared/exception/`
### Domain Services
- Return domain objects, not DTOs
- Contain business rules and state transitions
- Use domain events for cross-cutting side effects (publishing, notifications)
- Located in `domain/{submodule}/service/`
### Controllers
- Transport only: extract auth context, bind request params, wrap responses
- No business logic in controllers
- Located in `com.iflytek.skillhub.controller/`
### Query Repositories
- Handle read-model joins and presentation projection
- Return DTOs or presentation models
- Located in `com.iflytek.skillhub.repository/`
- Named like `*QueryRepository` (e.g., `GovernanceQueryRepository`, `MySkillQueryRepository`)
### App Services
- Workflow orchestration: coordinate domain services and query repositories
- Should express "what this endpoint does", not "how it assembles DTOs"
- Located in `com.iflytek.skillhub.service/`
### Logging
- Use SLF4J with structured logging
- Use MDC for request tracing
- Log at appropriate levels: INFO for business events, DEBUG for troubleshooting, ERROR for failures
## TypeScript / Frontend Conventions
### Type Safety
- Strict TypeScript mode. No `any` types.
- Use generated OpenAPI types from `web/src/api/generated/schema.d.ts` for all API interactions.
- Additional types in `web/src/types/`
### Data Fetching
- **Always use TanStack Query** (`@tanstack/react-query`) for server state
- **Never use `useEffect`** for data fetching
- Use `openapi-fetch` client for type-safe API calls
### Component Composition
- **Radix UI** primitives: `@radix-ui/react-dropdown-menu`, `@radix-ui/react-select`
- **class-variance-authority** (cva) for component variants
- **clsx** + **tailwind-merge** for class merging
- **`cn()` utility**: `web/src/shared/lib/utils.ts`
- shadcn/ui is NOT used as a library
### State Management
- **TanStack Query** for server state (API data, caching, invalidation)
- **Zustand** for local/UI state (theme, sidebar, modals, form state)
### Feature-Sliced Design
| Layer | Path | Purpose |
|-------|------|---------|
| Pages | `web/src/pages/` | Route-level page components |
| Features | `web/src/features/` | Self-contained business features |
| Entities | `web/src/entities/` | Domain entity display logic |
| Shared | `web/src/shared/` | Reusable UI components, hooks, utilities |
Place code at the lowest appropriate layer. Do not put page-level logic in shared.
### Styling
- Tailwind CSS for all styling
- Follow existing component patterns
- Use `cn()` for conditional class merging
### Internationalization
- Use i18next + react-i18next
- All user-facing text must be translatable
- Translation keys in `web/src/i18n/`
## Testing Philosophy
### Backend
- JUnit 5 + Mockito + AssertJ
- Use Spring Boot test slices where possible (`@WebMvcTest`, `@DataJpaTest`)
- Test behaviors, not implementations
- Use `make test-backend-app` (includes `-am` for dependent modules)
- Never run `./mvnw -pl skillhub-app clean test` directly — stale Maven cache causes misleading errors
### Frontend
- Vitest for unit tests
- Playwright for E2E tests
- Test component behavior and user interactions
## Common Pitfalls
- **Maven multi-module**: Always use `-am` flag or Makefile targets to include dependent modules
- **OpenAPI types**: Must regenerate and commit after API contract changes
- **String identity**: Never use numeric types for user identifiers
- **Controller business logic**: Move to domain service or app service
- **Complex read-models in app service**: Extract to query repository

View file

@ -0,0 +1,194 @@
---
name: dev-workflow
description: The complete development workflow for SkillHub contributors including local dev, staging validation, testing, and PR creation. Ensures agents follow the correct sequence of steps.
license: Apache-2.0
---
# Development Workflow Skill
## Trigger
Use this skill when:
- Starting local development
- Running tests or validation
- Preparing a pull request
- Setting up the development environment
- Working with parallel agent worktrees
## Prerequisites
- Java 21+ (`java -version`)
- Maven wrapper (`./mvnw` in `server/`)
- Node.js + pnpm
- Docker + docker compose
- `gh` CLI (for PR creation)
- `curl` (for smoke tests and health checks)
## Workflow Stages
### Stage 1: Local Development (fast iteration)
**One-command start:**
```bash
make dev-all # Start full stack: Postgres, Redis, MinIO, scanner, backend, frontend
make dev-all-down # Stop everything
make dev-all-reset # Full reset (clears data volumes)
make dev-status # Check service status
```
**Access points:**
- Web UI: `http://localhost:3000`
- Backend API: `http://localhost:8080`
- Scanner: `http://localhost:8000`
**Individual components:**
```bash
make dev # Start dependency services only (Postgres, Redis, MinIO, scanner)
make dev-server # Start backend in foreground (blocking)
make dev-web # Start Vite dev server (HMR enabled)
make dev-server-restart # Restart backend process
make dev-down # Stop dependency services
make dev-logs # View backend logs (use SERVICE=frontend for frontend logs)
```
**Backend development**: After editing Java code, run `make dev-server-restart`.
**Frontend development**: Vite HMR enabled — save a file for instant browser updates.
**Scanner**: The security scanner is enabled by default in dev. Health checked at `http://localhost:8000/health`.
### Stage 2: Testing
| Command | Scope | Notes |
|---------|-------|-------|
| `make test-backend-app` | Backend unit tests | skillhub-app + dependencies (`-am`) |
| `make test-backend` | All backend modules | All modules via `./mvnw test` |
| `make test-frontend` | Frontend unit tests | Vitest (pnpm run test) |
| `make test-e2e-frontend` | Frontend E2E tests | Playwright |
| `make test-e2e-smoke-frontend` | Frontend E2E smoke | Playwright subset |
| `make typecheck-web` | TypeScript type check | `tsc --noEmit` |
| `make lint-web` | ESLint check | Frontend linting |
**Important**: Never run `./mvnw -pl skillhub-app clean test` directly under `server/`.
Use `-am` or Makefile targets to include dependent modules.
### Stage 3: Staging Regression (pre-PR validation)
```bash
make staging # Build backend Docker image + frontend static + smoke test
make staging-down # Tear down
make staging-logs # View backend logs
SERVICE=web make staging-logs # View Nginx logs
```
Staging validates the containerized deployment path:
- Backend: built as Docker image from local source (`Dockerfile.dev`)
- Frontend: built as static files (`pnpm build`), served by Nginx
- Dependencies: same Postgres/Redis/MinIO as local dev
- Smoke test runs against staging via `scripts/smoke-test.sh`
**Staging URLs:**
- Web UI: `http://localhost`
- Backend API: `http://localhost:8080`
**Staging credentials** (for bootstrap admin):
- Username: `admin`
- Password: `Admin@staging2026`
### Stage 4: Pull Request
```bash
make pr # Push branch + create PR (requires gh CLI)
```
Requirements:
- `gh` CLI installed and authenticated
- Not on main/master branch
- All changes committed (will prompt if uncommitted changes exist)
### Useful Commands
| Command | Description |
|---------|-------------|
| `make generate-api` | Regenerate OpenAPI types from running backend |
| `make namespace-smoke` | Namespace workflow smoke test |
| `make db-reset` | Reset database only (Flyway migrate) |
| `make validate-release-config` | Validate release env vars (.env.release) |
| `./scripts/smoke-test.sh` | Basic API smoke test (health, auth, labels) |
| `./scripts/namespace-smoke-test.sh` | Namespace CRUD + membership smoke test |
| `./scripts/check-openapi-generated.sh` | Verify OpenAPI types are not stale |
| `make parallel-init TASK=name` | Create parallel worktree for agent |
### Mock Auth Users
| User ID | Role | Header |
|---------|------|--------|
| `local-user` | Regular user | `X-Mock-User-Id: local-user` |
| `local-admin` | Super admin | `X-Mock-User-Id: local-admin` |
Bootstrap admin (local profile):
- Username: `admin`
- Password: `ChangeMe!2026`
### Smoke Test Coverage
`scripts/smoke-test.sh` validates:
1. Health endpoint (`/actuator/health` → 200)
2. Prometheus metrics (`/actuator/prometheus` → 200)
3. Namespaces API (`/api/v1/namespaces` → 200)
4. Auth required (`/api/v1/auth/me` → 401 without session)
5. User registration flow (with CSRF)
6. Auth me with session
7. Password change
8. Logout + verify 401 after
9. Admin login
10. Label CRUD (admin only)
Additional smoke tests:
- `scripts/namespace-smoke-test.sh` — Namespace creation, membership, publishing
- `scripts/governance-smoke-test.sh` — Governance and moderation
- `scripts/promotion-smoke-test.sh` — Skill promotion between scopes
### Parallel Agent Workflow
For parallel agent development with isolated worktrees:
```bash
make parallel-init TASK=feature-name # Create worktree
make parallel-sync SOURCES="feat1 feat2" # Merge feature branches
make parallel-up SOURCES="feat1 feat2" # Merge + start dev environment
make parallel-down # Stop parallel environment
```
See `docs/13-parallel-workflow.md` for full details.
### Commit Style
Use conventional commit format:
```
<type>(<scope>): <description>
```
Examples:
```
feat(auth): add local account login
fix(publish): resolve null pointer when skill metadata is missing name
docs(deploy): clarify runtime image usage
test(namespace): add membership service edge case tests
refactor(review): extract query repository for governance list
chore(ci): add parallel workflow scripts
```
### Common Issues
| Issue | Solution |
|-------|----------|
| Backend won't start | Check Java version (`java -version`), must be 21+ |
| Port 8080 in use | `lsof -i :8080` to find and kill the process |
| Maven download timeout | Configure mirror in `~/.m2/settings.xml` |
| Frontend won't start | Run `make web-deps` to ensure node_modules exist |
| Staging build fails | Check `Dockerfile.dev` and ensure Maven build succeeds first |
| CSRF errors in tests | Ensure cookie jar is shared and CSRF token refreshed after login |

View file

@ -0,0 +1,124 @@
---
name: frontend-conventions
description: Coding conventions, architecture patterns, and testing rules for the SkillHub React frontend. Ensures agents follow Feature-Sliced Design and use the generated OpenAPI types.
license: Apache-2.0
---
# Frontend Conventions Skill
## Trigger
Use this skill when:
- Adding or modifying React/TypeScript frontend code
- Creating new pages, features, entities, or shared components
- Changing API client calls or data fetching patterns
## Rules
### Feature-Sliced Design
Place code at the lowest appropriate layer:
| Layer | Path | Purpose |
|-------|------|---------|
| Pages | `web/src/pages/` | Route-level page components |
| Features | `web/src/features/` | Business features (search, upload, review, etc.) |
| Entities | `web/src/entities/` | Domain entity display logic (skill, user, namespace) |
| Shared | `web/src/shared/` | Reusable UI components, hooks, utilities |
Current features:
- `admin` — Admin panel (user management, labels, search)
- `auth` — Login, OAuth flows, device auth
- `governance` — Skill governance actions (hide, yank, archive)
- `namespace` — Namespace management (members, settings)
- `notification` — User notifications and inbox
- `promotion` — Skill promotion between scopes
- `publish` — Skill upload/publish UI
- `report` — Skill reporting
- `review` — Review workflow UI
- `search` — Skill search and filtering
- `security-audit` — Security audit viewer
- `skill` — Skill detail, listing, cards
- `social` — Stars, ratings, subscriptions
- `token` — API token management
### Data Fetching
- **Always use TanStack Query** (`@tanstack/react-query`) for server state.
- **Never use `useEffect`** for data fetching.
- Use `openapi-fetch` client with generated types from `web/src/api/generated/schema.d.ts`.
- Never use `any` types.
### State Management
- **TanStack Query** for server state (API data, caching, invalidation, optimistic updates)
- **Zustand** for local/UI state (theme, sidebar, modals, form state)
### Component Composition
- **Radix UI** primitives: `@radix-ui/react-dropdown-menu`, `@radix-ui/react-select`
- **class-variance-authority** (cva) for component variants
- **clsx** + **tailwind-merge** for class merging
- **`cn()` utility**: `web/src/shared/lib/utils.ts`
- **shadcn/ui is NOT used** as a library
### API Type Generation
When backend OpenAPI contracts change:
```bash
make generate-api
```
This runs `openapi-typescript http://localhost:8080/v3/api-docs -o src/api/generated/schema.d.ts`.
Commit the updated `web/src/api/generated/schema.d.ts` with the PR.
To verify the generated file is not stale:
```bash
./scripts/check-openapi-generated.sh
```
### Styling
- **Tailwind CSS** for all styling
- **`cn()` utility** for conditional class merging
- Follow existing component patterns in `web/src/shared/components/`
### Internationalization
- **i18next** + **react-i18next** for translations
- All user-facing text must be translatable
- Translation keys in `web/src/i18n/`
### Build & Development
```bash
make dev-web # Start Vite dev server (HMR enabled)
make build-frontend # Production build
make typecheck-web # TypeScript type check (tsc --noEmit)
make lint-web # ESLint check
make test-frontend # Vitest unit tests
make test-e2e-frontend # Playwright E2E tests
make test-e2e-smoke-frontend # Playwright smoke tests
```
Vite HMR is enabled by default — save a file and the browser updates instantly.
### Frontend Dependencies
Key dependencies (from `web/package.json`):
- `react` 19, `react-dom` 19
- `@tanstack/react-query` 5
- `@tanstack/react-router` 1
- `@radix-ui/react-dropdown-menu`, `@radix-ui/react-select`
- `class-variance-authority`, `clsx`, `tailwind-merge`
- `openapi-fetch` 0.13
- `i18next`, `react-i18next`
- `zustand` 5
- `react-markdown`, `rehype-highlight`, `rehype-sanitize`
- `lucide-react` (icons)
- `sonner` (toasts)
Build tools: Vite 6, TypeScript 5.7, Vitest 3.2, Playwright 1.58

View file

@ -0,0 +1,93 @@
---
name: pr-submission
description: PR title format, commit conventions, and pre-PR checklist for SkillHub. Use when preparing or reviewing pull requests.
license: Apache-2.0
---
# PR Submission Skill
## Workflow
1. Identify the scope of your change (feature, bug fix, docs, test, refactor, chore)
2. Format PR title and commits using the conventions below
3. Run the pre-PR checklist commands
4. Open the PR with a descriptive body
## PR Title Format
Use conventional commit style:
```
<type>(<scope>): <description>
```
**Types:**
| Type | When to Use |
|------|-------------|
| `feat` | New feature or capability |
| `fix` | Bug fix |
| `docs` | Documentation changes only |
| `test` | Adding or updating tests |
| `refactor` | Code restructuring with no behavior change |
| `chore` | Build, CI, tooling, or maintenance tasks |
**Scopes:** Use module or domain names: `auth`, `search`, `publish`, `review`, `namespace`, `governance`, `deploy`, `ci`, `frontend`, `scanner`
**Examples:**
```
feat(auth): add local account login with password reset
fix(publish): resolve null pointer when skill metadata is missing name
docs(deploy): clarify runtime image usage
test(namespace): add membership service edge case tests
refactor(review): extract query repository for governance list
chore(ci): add parallel workflow scripts for multi-agent development
```
## Commit Message Format
Same convention as PR titles. One logical change per commit.
**Types:**
- **feat**: A new feature for the user
- **fix**: A bug fix for the user
- **docs**: Documentation changes only
- **test**: Adding or updating tests
- **refactor**: Code change that neither fixes a bug nor adds a feature
- **chore**: Changes to build process, CI, or maintenance tasks
**Examples:**
```
fix(auth): resolve session cookie conflict in device flow
feat(publish): support security scan before review submission
docs(skill-protocol): add nested SKILL.md discovery rules
test(search): verify jieba analysis with Chinese skill descriptions
refactor(storage): simplify LocalFile path normalization
```
## Pre-PR Checklist
- [ ] Backend tests pass: `make test-backend-app`
- [ ] Frontend typecheck passes: `make typecheck-web`
- [ ] If API changed: `make generate-api` was run and `web/src/api/generated/schema.d.ts` is committed
- [ ] Smoke test passes: `make staging`
- [ ] Follow existing module boundaries and dependency direction
- [ ] Add/update tests for new behavior
- [ ] Update design docs when APIs, auth flows, deployment, or operator workflows change
## PR Body Structure
When creating a PR, include:
1. **What** — Summary of the change
2. **Why** — Motivation (link to issue if applicable)
3. **How** — Key implementation details (especially for non-obvious decisions)
4. **Testing** — How to verify the change works
5. **Impact** — Breaking changes, migration notes, or rollout considerations
## Review Conventions
- When reviewing, cite the specific AGENTS.md rule that applies if suggesting a convention change
- For backend code, check dependency direction does not violate clean architecture rules
- For frontend code, check OpenAPI types are regenerated if API changed

View file

@ -0,0 +1,151 @@
---
name: skill-lifecycle
description: The authoritative skill lifecycle state model including container states, version states, review workflow states, visibility overlay, and governance actions. Ensures agents don't introduce invalid states or transitions.
license: Apache-2.0
---
# Skill Lifecycle Skill
## Trigger
Use this skill when:
- Modifying skill publish, review, or unpublish flows
- Adding or changing skill/version status fields
- Working on search, detail pages, or listing pages that show skill state
- Implementing governance actions (hide, yank, archive)
- Adding new state transitions or permission checks
## State Model
### Skill Container States
Enum `SkillStatus` (`domain/skill/SkillStatus.java`):
| Value | Meaning |
|-------|---------|
| `ACTIVE` | Skill is operational and can have versions published |
| `HIDDEN` | Skill hidden by platform governance (design doc says prefer boolean `hidden` flag instead) |
| `ARCHIVED` | Skill archived by owner/namespace admin, cannot publish new versions |
**Design-vs-code note**: `docs/14-skill-lifecycle.md` specifies `hidden` should be a governance
overlay (boolean flag) rather than a lifecycle enum state. The current code still defines
`SkillStatus.HIDDEN`. New code should use the `skill.hidden` boolean field, not the enum value.
### SkillVersion States
Enum `SkillVersionStatus` (`domain/skill/SkillVersionStatus.java`):
| Value | Meaning |
|-------|---------|
| `DRAFT` | Non-public draft, can resubmit or delete |
| `SCANNING` | Undergoing security scan |
| `SCAN_FAILED` | Security scan failed |
| `UPLOADED` | Uploaded but not yet submitted for review (or withdrawn from review) |
| `PENDING_REVIEW` | Frozen pending reviewer action |
| `PUBLISHED` | Currently distributable |
| `REJECTED` | Review denied, retained |
| `YANKED` | Was published, withdrawn from distribution |
### ReviewTask States
Enum `ReviewTaskStatus` (`domain/review/ReviewTaskStatus.java`):
| Value | Meaning |
|-------|---------|
| `PENDING` | Awaiting reviewer |
| `APPROVED` | Reviewer approved |
| `REJECTED` | Reviewer rejected |
### Visibility Model
Enum `SkillVisibility` (used in `SkillPublishService`):
| Value | Publish Path |
|-------|-------------|
| `PUBLIC` | Creates `PENDING_REVIEW` version, review task, security scan |
| `NAMESPACE_ONLY` | Same as PUBLIC but limited visibility scope |
| `PRIVATE` | Goes directly to `UPLOADED` status, no review task |
`SUPER_ADMIN` role bypasses review — versions go directly to `PUBLISHED`.
### Latest Version Pointer
`Skill.latestVersionId` is **only** the latest published pointer:
- Can only point to a `PUBLISHED` version
- May be `null` if no published version exists
- `latest` tag auto-follows this pointer (read-only)
- When yanking: recalculates to newest remaining `PUBLISHED` version, or `null`
### Key Transitions
| Action | From | To | Notes | Source |
|--------|------|-----|-------|--------|
| First upload (PUBLIC/NAMESPACE_ONLY) | — | `PENDING_REVIEW` | Review task created | `SkillPublishService` |
| First upload (SUPER_ADMIN) | — | `PUBLISHED` | Direct publish, `SkillPublishedEvent` emitted | `SkillPublishService` |
| First upload (PRIVATE) | — | `UPLOADED` | No review task, `latestVersionId` updated | `SkillPublishService` |
| Review approve | `PENDING_REVIEW` | `PUBLISHED` | Updates `latestVersionId` | Review workflow |
| Review reject | `PENDING_REVIEW` | `REJECTED` | Version retained | Review workflow |
| Withdraw review | `PENDING_REVIEW` | `UPLOADED` | Deletes pending `ReviewTask` | `SkillGovernanceService.withdrawPendingVersion` |
| Yank | `PUBLISHED` | `YANKED` | Recalculates `latestVersionId` | `SkillGovernanceService.yankVersion` |
| Hide | — | `hidden=true` | Independent overlay | `SkillGovernanceService.hideSkill` |
| Restore | — | `hidden=false` | Independent overlay | `SkillGovernanceService.unhideSkill` |
| Archive | `ACTIVE` | `ARCHIVED` | `SkillStatusChangedEvent` emitted | `SkillGovernanceService.archiveSkill` |
| Unarchive | `ARCHIVED` | `ACTIVE` | `SkillStatusChangedEvent` emitted | `SkillGovernanceService.unarchiveSkill` |
| New publish (existing pending) | `PENDING_REVIEW` | `UPLOADED` | Auto-withdraw + delete review task | `SkillPublishService` |
| Delete version | `DRAFT`/`REJECTED`/`SCAN_FAILED`/`UPLOADED` | — | Last version protected | `SkillGovernanceService.deleteVersion` |
### Yank Pointer Recalculation
When yanking the current `latestVersionId` (`SkillGovernanceService`):
1. Query all remaining `PUBLISHED` versions for the skill
2. Sort by `publishedAt` DESC, then `createdAt` DESC, then `id` DESC
3. Point `latestVersionId` to the top result, or `null` if none remain
### Lifecycle Projection
Read models (detail, my-skills, favorites, search) use `*QueryRepository` patterns:
- `headlineVersion` — Main display version for the page
- `publishedVersion` — Latest published version
- `ownerPreviewVersion` — Pending review version (visible to owner/namespace admin)
- `resolutionMode``PUBLISHED`, `OWNER_PREVIEW`, or `NONE`
**Public browsing, install, download, search only use `publishedVersion`.**
### Permission Boundaries
| Action | Who |
|--------|-----|
| Withdraw review | Submitter only |
| Delete version | Owner or namespace admin, only `DRAFT`/`REJECTED`/`SCAN_FAILED`/`UPLOADED` |
| Archive/unarchive | Owner or namespace admin (`ADMIN` or `OWNER` role) |
| Hide/restore | Platform governance (no permission check in code) |
| Yank | Platform governance (no permission check in code) |
| Publish PUBLIC skill | Namespace member (or `SUPER_ADMIN`) |
| Publish PRIVATE skill | Namespace member (or `SUPER_ADMIN`) |
### Delete Version Constraints
`SkillGovernanceService.deleteVersion` enforces:
- Only `DRAFT`, `REJECTED`, `SCAN_FAILED`, or `UPLOADED` versions can be deleted
- Cannot delete the last remaining version of a skill
- Deletes associated storage keys (individual files + `bundle.zip`)
- Deletes associated security scan records
- Updates `latestVersionId` if the deleted version was the pointer
- Storage deletion happens after transaction commit with compensation recording
### Domain Events
| Event | When Emitted |
|-------|-------------|
| `SkillStatusChangedEvent` | Archive or unarchive |
| `SkillPublishedEvent` | SUPER_ADMIN direct publish |
| `SkillVersionYankedEvent` | Yank action |
| `ReviewSubmittedEvent` | Create review task for PUBLIC/NAMESPACE_ONLY |
### Common Pitfalls
- Setting `SkillStatus.HIDDEN` directly — use `skill.setHidden(true)` via `SkillGovernanceService` instead
- Forgetting to recalculate `latestVersionId` after yank or version deletion
- Not auto-withdrawing pending versions when publishing a new version
- Missing the `confirmWarnings` two-step publish flow (warnings require explicit confirmation)
- Assuming all publish flows create review tasks — `PRIVATE` visibility skips review

View file

@ -0,0 +1,117 @@
---
name: testing-and-ci
description: Testing conventions, CI pipeline rules, and smoke test coverage for SkillHub. Ensures agents write tests correctly and understand the CI gate requirements.
license: Apache-2.0
---
# Testing and CI Skill
## Trigger
Use this skill when:
- Adding or modifying backend tests
- Adding or modifying frontend tests
- Changing CI/CD workflows
- Adding smoke tests or E2E tests
## Rules
### Backend Testing
Tests live alongside source in each module's `src/test/java/`:
- `server/skillhub-app/src/test/java/` — Controller integration tests, service tests
- `server/skillhub-domain/src/test/java/` — Domain service unit tests
- `server/skillhub-auth/src/test/java/` — Auth flow tests
**Tools**: JUnit 5 + Mockito + AssertJ + Spring Boot test slices (`@WebMvcTest`, `@DataJpaTest`)
**Build commands:**
```bash
make test-backend-app # skillhub-app + dependencies (includes -am)
make test-backend # all backend modules
```
**Never** run `./mvnw -pl skillhub-app clean test` directly under `server/`.
`skillhub-app` depends on sibling modules, and a standalone clean build can fall back to stale
artifacts from the local Maven repository, surfacing misleading `cannot find symbol` and
signature-mismatch errors. Use `-am`, or the Makefile targets above.
**Test naming conventions:**
- Controller tests: `{ControllerName}Test.java` (e.g., `SkillControllerTest.java`)
- Service tests: `{ServiceName}Test.java`
- Integration tests: `{FlowName}IntegrationTest.java`
- Security tests: `{ControllerName}SecurityTest.java`
### Frontend Testing
**Tools**: Vitest (unit), Playwright (E2E)
```bash
make test-frontend # Vitest unit tests (pnpm run test)
make test-e2e-frontend # Playwright E2E tests
make test-e2e-smoke-frontend # Playwright smoke tests (subset)
```
E2E tests live in `web/e2e/`.
### Smoke Tests
Smoke tests validate end-to-end operator workflows against a running backend:
| Script | Purpose |
|--------|---------|
| `scripts/smoke-test.sh` | Basic API health, auth, label CRUD |
| `scripts/namespace-smoke-test.sh` | Namespace creation, membership, publishing |
| `scripts/governance-smoke-test.sh` | Governance and moderation flows |
| `scripts/promotion-smoke-test.sh` | Skill promotion between scopes |
When operator-facing workflows change, update the corresponding smoke test.
### CI Pipeline
GitHub Actions workflows in `.github/workflows/`:
| Workflow | Trigger | Purpose |
|----------|---------|---------|
| `pr-tests.yml` | PR | Backend + frontend unit tests |
| `pr-e2e.yml` | PR | E2E smoke tests against staging |
| `pr-batch-test-deploy.yml` | workflow_dispatch | Batch test and deploy |
| `publish-images.yml` | release published / workflow_dispatch | Build and publish Docker images to GHCR |
| `deploy-docs.yml` | push to docs | Deploy documentation site |
| `issue-triage.yml` | issues | Auto-triage incoming issues |
| `issue-backlog-rescore.yml` | cron (every 6h) | Rescore backlog issues |
| `release-notes.yml` | workflow_dispatch | Generate release notes |
| `deepwiki.yml` | release published | Update DeepWiki documentation |
| `claim-issue-reward.yml` | issue_comment | Auto-claim issue rewards |
| `statistic-member-reward.yml` | cron/schedule | Calculate member rewards |
All workflows live in `.github/workflows/`. Deno scripts for triage, release notes, and rewards
live in `.github/scripts/`.
### Staging
Before opening a PR, validate with staging:
```bash
make staging # Build backend Docker image + frontend static + smoke test
make staging-down # Tear down
SERVICE=web make staging-logs # View Nginx logs
```
Staging validates the containerized deployment path:
- Backend: built as Docker image from local source (`Dockerfile.dev`)
- Frontend: built as static files (`pnpm build`), served by Nginx
- Dependencies: same Postgres/Redis/MinIO as local dev
If staging passes, the environment stays running at:
- Web UI: `http://localhost`
- Backend API: `http://localhost:8080`
### Pre-PR Testing Checklist
- [ ] `make test-backend-app` passes
- [ ] `make typecheck-web` passes
- [ ] `make lint-web` passes (if frontend changed)
- [ ] `make staging` passes (full regression)
- [ ] If API changed: `make generate-api` run and generated file committed
- [ ] New behavior has corresponding tests

1
.gitignore vendored
View file

@ -82,7 +82,6 @@ docs/review/
docs/superpowers/
# Local workspace metadata
AGENTS.md
CLAUDE.md
# Local config file

599
AGENTS.md Normal file
View file

@ -0,0 +1,599 @@
# SkillHub — AGENTS.md
**SkillHub** is an **enterprise-grade, self-hosted agent skill registry** for publishing,
discovering, and managing reusable skill packages across an organization. It provides a **REST API
backend**, a **React web UI**, a **security scanner**, and a **ClawHub CLI compatibility layer**.
## Quick Reference
| Item | Value |
|------------|------------------------------------------------------------|
| Backend | Spring Boot 3.2.3, Java 21, Maven multi-module (7 modules) |
| Frontend | React 19, TypeScript, Vite, pnpm |
| Scanner | Python (FastAPI), port 8000 |
| Database | PostgreSQL 16 (Flyway migrations) |
| Cache | Redis 7 (sessions, distributed locks, idempotency) |
| Storage | LocalFile (dev) / S3/MinIO (prod) |
| Build | `make dev-all` (dev), `make staging` (pre-PR) |
| Docs | `docs/` (design), `document/` (VitePress user guide) |
| CI | GitHub Actions (`.github/workflows/`) |
## Directory Map
```
skillhub/
├── server/ # Maven multi-module Spring Boot backend
│ ├── skillhub-app/ # Application layer: bootstrap, controllers, assembly
│ │ ├── bootstrap/ # Bootstrap admin & local dev data initializers
│ │ ├── compat/ # ClawHub CLI compatibility layer controllers
│ │ ├── config/ # Spring configuration classes
│ │ ├── controller/ # REST controllers (transport only)
│ │ │ ├── admin/ # Admin controllers (user mgmt, labels, search)
│ │ │ ├── portal/ # Portal controllers (skills, governance, security)
│ │ │ └── support/ # Package extractors (zip, multipart)
│ │ ├── dto/ # Request/response DTOs
│ │ ├── exception/ # Exception handling
│ │ ├── filter/ # Servlet filters (auth context, rate limiting)
│ │ ├── listener/ # Event listeners (notification recipients, etc.)
│ │ ├── metrics/ # Micrometer metrics
│ │ ├── projection/ # Lifecycle projection models
│ │ ├── ratelimit/ # Rate limiting logic
│ │ ├── repository/ # Query repositories (read-model assembly)
│ │ ├── security/ # Security configuration
│ │ ├── service/ # App services (workflow orchestration)
│ │ ├── stream/ # SSE streaming endpoints
│ │ ├── task/ # Background task scheduling
│ │ └── SkillhubApplication.java # Spring Boot entry point
│ │
│ ├── skillhub-domain/ # Domain layer: entities, rules, services (innermost)
│ │ ├── audit/ # AuditLog entity, repository, service
│ │ ├── auth/ # Password reset entities
│ │ ├── event/ # Domain event classes (SkillPublishedEvent, etc.)
│ │ ├── governance/ # Governance notification service
│ │ ├── idempotency/ # Idempotency records
│ │ ├── label/ # Skill label management
│ │ ├── namespace/ # Namespace, members, roles, policies
│ │ ├── report/ # Skill reporting/governance
│ │ ├── review/ # Review tasks, promotion requests
│ │ ├── security/ # Security scanning domain model
│ │ ├── shared/ # Shared domain utilities
│ │ │ └── exception/ # Domain exceptions (LocalizedDomainException, etc.)
│ │ ├── skill/ # Core skill entities and services
│ │ │ ├── metadata/ # SKILL.md frontmatter parsing
│ │ │ ├── service/ # Skill domain services (publish, query, governance)
│ │ │ └── validation/ # Package validation (SkillPackagePolicy, etc.)
│ │ ├── social/ # Star, rating, subscription entities
│ │ └── user/ # UserAccount, profile moderation
│ │
│ ├── skillhub-auth/ # Authentication & authorization
│ │ ├── config/ # Spring Security configuration
│ │ ├── device/ # OAuth Device Flow for CLI auth
│ │ ├── identity/ # Identity binding service
│ │ ├── local/ # Local (password) auth
│ │ ├── merge/ # Account merging
│ │ ├── oauth/ # OAuth2 login handlers
│ │ ├── policy/ # Route security policies
│ │ ├── rbac/ # RBAC service and role definitions
│ │ ├── token/ # API token management
│ │ └── user/ # User-related auth services
│ │
│ ├── skillhub-search/ # Search SPI + PostgreSQL full-text implementation
│ │ ├── postgres/ # PostgresFullTextIndexService, QueryService
│ │ └── service/ # Search SPI interfaces
│ │
│ ├── skillhub-storage/ # Object storage SPI
│ │ ├── local/ # LocalFileStorageService
│ │ └── s3/ # S3StorageService (AWS SDK v2)
│ │
│ ├── skillhub-infra/ # Infrastructure: JPA repos, utilities
│ │ └── repository/ # Spring Data JPA repository implementations
│ │
│ ├── skillhub-notification/ # Notification service (SSE, email)
│ │ ├── domain/ # Notification domain model
│ │ ├── service/ # Notification delivery services
│ │ └── sse/ # SSE endpoint support
│ │
│ ├── Dockerfile.dev # Dockerfile for staging builds
│ ├── Dockerfile # Production multi-stage build
│ ├── pom.xml # Parent POM (Spring Boot 3.2.3 parent)
│ └── scripts/
│ └── run-dev-app.sh # Local dev startup script
├── web/ # React frontend (Vite + pnpm)
│ ├── src/
│ │ ├── api/ # OpenAPI-generated types + fetch client
│ │ │ └── generated/
│ │ │ └── schema.d.ts # Generated OpenAPI types (CHECKED IN)
│ │ ├── app/ # Router, layout, global providers
│ │ ├── docs/ # In-app documentation pages
│ │ ├── entities/ # Domain entity display logic
│ │ │ ├── skill/ # Skill card, detail components
│ │ │ ├── user/ # User profile components
│ │ │ └── namespace/ # Namespace display components
│ │ ├── features/ # Business feature modules
│ │ │ ├── admin/ # Admin panel features
│ │ │ ├── auth/ # Login, OAuth flows
│ │ │ ├── governance/ # Skill governance actions
│ │ │ ├── namespace/ # Namespace management
│ │ │ ├── notification/ # User notifications
│ │ │ ├── promotion/ # Skill promotion workflows
│ │ │ ├── publish/ # Skill upload/publish UI
│ │ │ ├── report/ # Skill reporting
│ │ │ ├── review/ # Review workflow UI
│ │ │ ├── search/ # Skill search and filtering
│ │ │ ├── security-audit/ # Security audit viewer
│ │ │ ├── skill/ # Skill detail, listing
│ │ │ ├── social/ # Stars, ratings, subscriptions
│ │ │ └── token/ # API token management
│ │ ├── i18n/ # Internationalization
│ │ ├── pages/ # Route-level page components
│ │ ├── shared/ # Shared UI, hooks, utilities
│ │ │ ├── components/ # Reusable UI components
│ │ │ ├── hooks/ # Custom React hooks
│ │ │ ├── lib/
│ │ │ │ └── utils.ts # cn() class merging utility
│ │ │ └── ui/ # Radix UI-based primitives
│ │ └── types/ # Additional TypeScript types
│ ├── e2e/ # Playwright E2E tests
│ ├── nginx.conf.template # Nginx runtime config template
│ ├── Dockerfile # Multi-stage build (Node → Nginx)
│ └── package.json # Dependencies (React 19, TanStack Query, Radix UI, etc.)
├── scanner/ # Security scanner (Python/FastAPI)
│ ├── docs/ # Scanner documentation
│ ├── examples/ # Example scan inputs/outputs
│ └── Dockerfile # Scanner container build
├── docs/ # Design documents (source of truth)
│ ├── prds/ # Product requirement documents
│ ├── skillhub/ # VitePress user guide source
│ └── superpowers/ # Internal tooling docs
├── document/ # VitePress documentation site (published)
│ ├── docs/ # Markdown documentation
│ ├── src/ # VitePress theme
│ └── i18n/ # Internationalization
├── deploy/k8s/ # Kubernetes manifests (basic)
├── monitoring/ # Prometheus + Grafana stack
├── scripts/ # Build, test, and deployment scripts
│ ├── smoke-test.sh # Basic API smoke test
│ ├── namespace-smoke-test.sh # Namespace workflow smoke test
│ ├── governance-smoke-test.sh # Governance flow smoke test
│ ├── promotion-smoke-test.sh # Promotion flow smoke test
│ ├── check-openapi-generated.sh # Verify OpenAPI SDK is not stale
│ ├── validate-release-config.sh # Validate release env configuration
│ ├── dev-process.sh # Local process manager (PID-based)
│ ├── runtime.sh # Runtime deployment script
│ ├── parallel-init.sh # Parallel worktree initialization
│ ├── parallel-sync.sh # Merge worktrees in integration branch
│ ├── parallel-up.sh # Merge + start dev environment
│ ├── parallel-down.sh # Stop parallel dev environment
│ └── prepare-pr-batch.sh # Batch PR preparation
├── .github/
│ ├── workflows/ # GitHub Actions CI/CD
│ ├── ISSUE_TEMPLATE/ # Issue templates
│ └── scripts/ # Deno scripts for triage, release notes, rewards
├── AGENTS.md # AI agent rules (this file)
├── .agents/skills/ # Focused AI skill definitions
├── Makefile # Top-level build/test/dev orchestration
├── docker-compose.yml # Local dev dependency services
├── compose.release.yml # Production release compose file
├── CONTRIBUTING.md # Contribution guidelines
├── CODE_OF_CONDUCT.md # Community standards
└── README.md # Project overview
```
**Key Locations for Common Tasks:**
| Task | Where to Look |
|------|---------------|
| Add REST endpoint | `server/skillhub-app/src/main/java/com/iflytek/skillhub/controller/` |
| Add domain entity/service | `server/skillhub-domain/src/main/java/com/iflytek/skillhub/domain/` |
| Add auth logic | `server/skillhub-auth/src/main/java/com/iflytek/skillhub/auth/` |
| Add search logic | `server/skillhub-search/src/main/java/com/iflytek/skillhub/search/` |
| Add query repository | `server/skillhub-app/src/main/java/com/iflytek/skillhub/repository/` |
| Change RBAC/roles | `server/skillhub-auth/src/main/java/com/iflytek/skillhub/auth/rbac/` |
| Change skill validation | `server/skillhub-domain/src/main/java/com/iflytek/skillhub/domain/skill/validation/` |
| Add frontend page | `web/src/pages/` |
| Add frontend feature | `web/src/features/` |
| Add shared component | `web/src/shared/components/` |
| Change API contract | Backend controller → run `make generate-api` → commit generated file |
| Add smoke test | `scripts/` (new `.sh` file) |
| Add E2E test | `web/e2e/` (Playwright) |
| Add backend test | `server/skillhub-*/src/test/java/` (alongside source module) |
## Critical Rules
### Do Not Manually Edit Generated Files
- `web/src/api/generated/schema.d.ts` — regenerated via `make generate-api`
- `document/docs/` — auto-generated user documentation (VitePress)
- `server/skillhub-app/src/main/java/com/iflytek/skillhub/dto/` — some DTOs may be generated
### After Making Changes
**Backend changes:**
- Edit Java code → `make dev-server-restart` (local dev)
- Add/modify controller → `make generate-api` to regenerate frontend types
- Add/modify domain service → `make test-backend-app` to verify tests
**Frontend changes:**
- Edit TypeScript/React → Vite HMR handles reload automatically
- After `make generate-api` → commit updated `web/src/api/generated/schema.d.ts`
**Always run before PR:**
```bash
make test-backend-app # Backend tests (with dependent modules)
make typecheck-web # Frontend type check
make lint-web # Frontend lint
make staging # Full staging regression + smoke test
```
### File-Specific Requirements
- **Controllers** (`skillhub-app/controller/`) are transport only: extract auth context,
bind request params, wrap responses. No business logic.
- **App Services** (`skillhub-app/service/`) orchestrate workflows. Do not embed complex
read-model assembly here — extract to query repositories.
- **Query Repositories** (`skillhub-app/repository/`) handle read-model joins and presentation
projection. Named like `*QueryRepository`.
- **Domain Services** (`skillhub-domain/*/service/`) contain business rules and state transitions.
Return domain objects, not DTOs.
- **Repository Interfaces** are defined in `skillhub-domain`, implemented in `skillhub-infra`.
- **Domain Exceptions** use `LocalizedDomainException` for user-facing messages with i18n keys.
- **Package-info files** (`package-info.java`) should exist for all packages.
## Development Workflow
### Build & Start
```bash
make dev-all # Start full stack: Postgres, Redis, MinIO, backend, frontend
make dev-all-down # Stop everything
make dev-all-reset # Full reset (clears data volumes)
make dev-status # Check service status
make dev-server-restart # Restart backend after Java changes
```
**Access points:**
- Web UI: `http://localhost:3000`
- Backend API: `http://localhost:8080`
- Scanner: `http://localhost:8000`
**Local mock users** (no password needed):
| User ID | Role | Header |
|---------|------|--------|
| `local-user` | Regular user | `X-Mock-User-Id: local-user` |
| `local-admin` | Super admin | `X-Mock-User-Id: local-admin` |
**Bootstrap admin** (password-based, local profile):
- Username: `admin` / Password: `ChangeMe!2026`
- Disable with `BOOTSTRAP_ADMIN_ENABLED=false`
### Lint & Format
```bash
# Backend: enforced by Maven build (no separate lint target)
# Frontend:
make lint-web # ESLint check
make typecheck-web # TypeScript check
```
### Testing
```bash
make test-backend-app # Backend unit tests (skillhub-app + dependencies)
make test-backend # All backend module tests
make test-frontend # Frontend unit tests (Vitest)
make test-e2e-frontend # Frontend E2E tests (Playwright)
make test-e2e-smoke-frontend # Frontend E2E smoke tests
./scripts/smoke-test.sh # API smoke test
make namespace-smoke # Namespace workflow smoke test
```
### Staging (Pre-PR Regression)
```bash
make staging # Build backend Docker image + frontend static + smoke test
make staging-down # Tear down
SERVICE=web make staging-logs # View Nginx logs
```
Staging validates the containerized deployment path:
- Backend: built as Docker image from local source
- Frontend: built as static files, served by Nginx
- Dependencies: same Postgres/Redis/MinIO as local dev
### Parallel Agent Workflow
For parallel development with isolated worktrees:
```bash
make parallel-init TASK=feature-name
```
Creates dedicated Claude, Codex, and integration worktrees as sibling directories.
See `docs/13-parallel-workflow.md` for details.
## PR Submission
### PR Title Format
Use conventional commit style:
```
<type>(<scope>): <description>
```
**Types:**
- `feat`: New feature
- `fix`: Bug fix
- `docs`: Documentation changes
- `test`: Adding or updating tests
- `refactor`: Code restructuring (no behavior change)
- `chore`: Build, CI, or maintenance tasks
**Scopes:** Module or domain name (e.g., `auth`, `search`, `publish`, `review`, `namespace`)
**Examples:**
```
feat(auth): add local account login
fix(publish): resolve null pointer in skill validation
docs(deploy): clarify runtime image usage
test(namespace): add membership service tests
refactor(review): extract query repository for governance list
chore(ci): add parallel workflow scripts
```
### Pre-PR Checklist
- [ ] Backend tests pass: `make test-backend-app`
- [ ] Frontend typecheck passes: `make typecheck-web`
- [ ] If API changed: `make generate-api` was run and `web/src/api/generated/schema.d.ts` is committed
- [ ] Smoke test passes: `make staging`
- [ ] Follow existing module boundaries and dependency direction
- [ ] Add/update tests for new behavior
- [ ] Update docs when APIs, auth flows, deployment, or operator workflows change
## Core Concepts
### Backend Clean Architecture
```
app → domain, auth, search, storage, infra, notification
infra → domain # implements domain repository interfaces
auth → domain
search → domain
notification → domain
storage → (independent) # pure SPI
```
**Design intent**: `skillhub-domain` is the innermost layer. It defines entities, repository
interfaces, and domain services without depending on infra, auth, search, or storage.
**Code reality**: `skillhub-domain` declares a Maven dependency on `skillhub-storage`, and several
domain services (`SkillHardDeleteService`, `SkillDownloadService`, `SkillPublishService`,
`SkillGovernanceService`, `SkillQueryService`, `SkillStorageDeletionCompensationService`) import
`com.iflytek.skillhub.storage.ObjectStorageService`. This is an existing deviation from the ideal.
New code should avoid adding further cross-module dependencies from domain.
### Repository / Query Boundary
When adding new read logic, follow these rules:
1. **Domain repository ports** (`skillhub-domain`): Aggregate reads, state transitions, rule
evaluation. Used by domain services.
2. **App query repositories** (`com.iflytek.skillhub.repository`): Read-model assembly that joins
multiple sources, presentation projection. Used by controllers and app services.
3. **App services** (`com.iflytek.skillhub.service`): Workflow orchestration. Should express "what
this endpoint does", not "how it assembles DTOs".
4. **Direct SQL / EntityManager**: Only when necessary, with class-level comment explaining why.
**Do not** add complex read-model assembly logic inside app services. Extract it into a query
repository when it joins multiple sources, does presentation projection, or is reused across services.
### Skill Lifecycle
`SkillVersionStatus` values: `DRAFT`, `SCANNING`, `SCAN_FAILED`, `UPLOADED`, `PENDING_REVIEW`,
`PUBLISHED`, `REJECTED`, `YANKED`.
`SkillStatus` enum values: `ACTIVE`, `HIDDEN`, `ARCHIVED`.
The design doc (`docs/14-skill-lifecycle.md`) specifies that `hidden` should be treated as a
governance overlay rather than a lifecycle state. The current code still defines
`SkillStatus.HIDDEN` in the enum. Follow the design doc's intent for new code.
**Key transitions:**
- Normal user first upload → `PENDING_REVIEW` (no initial DRAFT)
- SUPER_ADMIN first upload → `PUBLISHED` (direct publish)
- Review approve → `PENDING_REVIEW``PUBLISHED` (updates `latestVersionId`)
- Review reject → `PENDING_REVIEW``REJECTED`
- Withdraw review → `PENDING_REVIEW``UPLOADED` (also deletes pending review_task)
- Yank → `PUBLISHED``YANKED` (must recalculate `latestVersionId`)
- Hide/restore → independent `hidden` flag (governance overlay)
- Archive/Unarchive → `ACTIVE``ARCHIVED` (container state)
### Namespace Coordinate System
SkillHub uses `@{namespace_slug}/{skill_slug}`:
- `@global/my-skill` — Platform-level public namespace
- `@my-team/my-skill` — Team/department namespace
ClawHub CLI compatibility maps:
| SkillHub | Canonical Slug |
|----------|---------------|
| `@global/my-skill` | `my-skill` |
| `@team-name/my-skill` | `team-name--my-skill` |
### Authentication
- Web: OAuth2 (GitHub) + local password auth
- CLI: OAuth Device Flow (web authorization → CLI credentials)
- Programmatic: API tokens (prefix-based secure hashing)
- Session: Spring Session + Redis
### RBAC
Platform roles: `SUPER_ADMIN`, `SKILL_ADMIN`, `USER_ADMIN`, `AUDITOR`
Namespace roles: `OWNER`, `ADMIN`, `MEMBER`
### Skill Package Protocol
- Root: `SKILL.md` with YAML frontmatter (`name`, `description` required)
- Allowed extensions (50+ types): `.md`, `.txt`, `.json`, `.yaml`, `.yml`, `.js`, `.ts`, `.py`,
`.sh`, `.png`, `.jpg`, `.svg`, and many more (see `SkillPackagePolicy.ALLOWED_EXTENSIONS`)
- Limits: 10MB per file, 100MB total, 500 files max
- File type signatures validated (PNG magic bytes, SVG content check, etc.)
### Frontend State Management
- **TanStack Query** (`@tanstack/react-query`): All server state (API data)
- **Zustand**: Local/UI state (theme, sidebar, modals)
- **Never** use `useEffect` for data fetching
### Frontend Component Composition
- **Radix UI** primitives: `@radix-ui/react-dropdown-menu`, `@radix-ui/react-select`
- **class-variance-authority** (cva) for component variants
- **clsx** + **tailwind-merge** for class merging
- **`cn()` utility**: `web/src/shared/lib/utils.ts`
- shadcn/ui is NOT used as a library — only Radix primitives + utility composition
## Common Patterns
### Code Style
**Java:**
- User identity type is always `String` throughout the codebase
- Use Java 21 features (records, pattern matching, virtual threads)
- Follow existing naming patterns in the domain layer
- Error strings for `DomainBadRequestException`, etc. should be clear and actionable
**TypeScript:**
- Strict mode. No `any` types.
- Use generated OpenAPI types for all API interactions.
- Feature-Sliced Design: place code at the lowest appropriate layer.
### Testing Philosophy
- Backend: JUnit 5 + Mockito + AssertJ
- Frontend: Vitest for unit tests, Playwright for E2E
- **Use `make test-backend-app`** (includes `-am` for dependent modules) — never run
`./mvnw -pl skillhub-app clean test` directly, as it can use stale Maven cache artifacts
- Test behaviors, not implementations
- Use Spring Boot test slices where possible (`@WebMvcTest`, `@DataJpaTest`)
### Frontend Testing
```bash
make test-frontend # Vitest unit tests
make test-e2e-frontend # Playwright E2E
make test-e2e-smoke-frontend # Playwright smoke (subset of E2E)
```
### Logging Conventions
- **Backend**: SLF4J + Spring Boot logging. Use structured logging with MDC for request tracing.
- **Frontend**: `console.error` for errors, `console.warn` for deprecations, avoid `console.log` in production code.
- **Scanner**: Python logging module with structured JSON output.
### Security
- API tokens are stored as prefix-based secure hashes, never in plaintext
- OAuth2 client secrets and other secrets must not be logged or committed
- User identity is `String` (supports external SSO/OIDC/SCIM identity sources)
- The bootstrap admin (`BOOTSTRAP_ADMIN_ENABLED`) is for zero-config quickstart only
## Search Tips
```bash
# Find all REST endpoints
rg "@(Get|Post|Put|Delete|Patch)Mapping" --type java
# Find domain services
rg "class.*Service" server/skillhub-domain/
# Find query repositories
rg "QueryRepository" server/skillhub-app/
# Find controllers
rg "@RestController" server/skillhub-app/
# Find RBAC role checks
rg "@PreAuthorize" server/skillhub-app/
# Find skill validation logic
rg "SkillPackage" server/skillhub-domain/
# Find frontend features
rg "export" web/src/features/
# Find OpenAPI type generation script
rg "generate-api" web/package.json
# Find event listeners
rg "@EventListener" server/
```
## Design Philosophy
- **Hub first**: The server-side registry is the core product; CLI and agent integrations are entry capabilities
- **Compatibility first**: Support `SKILL.md` format and common directory conventions
- **Layered architecture**: Search and object storage must have replaceable boundaries (SPI pattern)
- **Open authentication**: OAuth2-based, extensible to multiple providers beyond GitHub
- **Audit first**: Enterprise distribution requires audit trails for publish, download, delete, and authorization
## References
### Essential Files
- **`Makefile`** — All build/test/dev automation targets
- **`CONTRIBUTING.md`** — Contribution guidelines and commit style
- **`CODE_OF_CONDUCT.md`** — Community standards
- **`server/pom.xml`** — Maven parent POM, module definitions, dependency versions
- **`web/package.json`** — Frontend dependencies and scripts
- **`.github/workflows/pr-tests.yml`** — PR test pipeline
- **`.github/workflows/publish-images.yml`** — Docker image publish to GHCR
### Key Directories
- **`server/skillhub-domain/`** — Core domain (entities, services, rules)
- **`server/skillhub-app/controller/`** — REST API endpoints
- **`server/skillhub-app/repository/`** — Query repositories
- **`server/skillhub-app/compat/`** — ClawHub CLI compatibility layer
- **`server/skillhub-auth/`** — Authentication and authorization
- **`web/src/features/`** — Frontend feature modules
- **`web/src/api/generated/`** — Generated OpenAPI types
- **`docs/`** — Design documents
- **`scripts/`** — Build, test, and deployment scripts
### Important Scripts
- **`scripts/smoke-test.sh`** — Basic API smoke test
- **`scripts/namespace-smoke-test.sh`** — Namespace workflow test
- **`scripts/check-openapi-generated.sh`** — Verify frontend SDK is current
- **`scripts/validate-release-config.sh`** — Validate production env config
- **`scripts/dev-process.sh`** — Local process manager (PID-based lifecycle)
- **`scripts/parallel-init.sh`** — Create isolated worktrees for parallel development
### Design Documents
- **`00-product-direction.md`** — Product positioning, MVP scope, coordinate system
- **`01-system-architecture.md`** — System architecture, module structure, dependency rules
- **`02-domain-model.md`** — Domain entities and relationships
- **`03-authentication-design.md`** — OAuth2, CLI Device Flow, API tokens
- **`04-search-architecture.md`** — Search SPI and implementations
- **`05-business-flows.md`** — Business process flows
- **`06-api-design.md`** — API contract specifications
- **`07-skill-protocol.md`** — SKILL.md format, package structure, CLI compatibility
- **`08-frontend-architecture.md`** — Frontend patterns and conventions
- **`14-skill-lifecycle.md`** — Skill state model (authoritative)
- **`dev-workflow.md`** — Local development workflow guide
### External Resources
- **SkillHub Docs**: https://iflytek.github.io/skillhub/
- **DeepWiki**: https://deepwiki.com/iflytek/skillhub
- **Discord**: https://discord.gg/qHYvtDNPHS
- **OpenSkills**: https://agents.md/ (skill package format reference)
- **OpenClaw**: https://github.com/openclaw/openclaw (CLI compatibility)
- **AstronClaw**: https://agent.xfyun.cn/astron-claw (cloud AI assistant integration)

View file

@ -6,6 +6,10 @@ SkillHub is a self-hosted registry for agent skills. Contributions should
preserve the existing architecture and product direction documented in
[`docs/`](./docs).
AI coding agents working in this repository should follow the rules in
[`AGENTS.md`](./AGENTS.md), which documents repository architecture,
dependency rules, and agent-specific conventions.
## Before You Start
- Read [`README.md`](./README.md) for local development commands.