mirror of
https://github.com/iflytek/skillhub.git
synced 2026-08-28 11:25:00 +00:00
ee33505b7c
* feat(security): extend scanner config with full analyzer options
Integrate skill-scanner's 8 analysis engines and policy configuration
into SkillHub's config system. Operators can now control behavioral,
LLM, Meta, AI Defense, VirusTotal, and trigger analyzers via
application.yml or environment variables.
Changes:
- Add Analyzers and Policy nested classes to SkillScannerProperties
- Create ScanOptions record to encapsulate analyzer flags
- Update SkillScannerService to pass options in /scan body and /scan-upload query params
- Wire ScanOptions through SkillScannerConfig and SkillScannerAdapter
- Extend application.yml with full scanner config block and env var overrides
- Update all tests to verify new configuration flow
All tests pass.
* feat(security): add domain model and integrate scan into publish flow
Add SCANNING/SCAN_FAILED status to SkillVersionStatus. Introduce
SecurityScanService, SecurityScanner port, ScanTask, SecurityAudit
and related domain types. Wire scan trigger into SkillPublishService
so non-auto-publish versions enter scanning when scanner is enabled,
falling back to review task creation when disabled.
* feat(security): add infra layer for scanner HTTP client and adapters
Add WebClient-based HttpClient abstraction with WebClientHttpClient
implementation. Add SkillScannerApiResponse record, SecurityScanException,
and SecurityAuditJpaRepository. Add webflux and test dependencies to
infra module.
* feat(security): add Redis stream consumers, audit API, and DB migration
Add AbstractStreamConsumer base class, ScanTaskConsumer for processing
scan results from Redis stream, and RedisScanTaskProducer. Add
RedisStreamConfig for stream/group initialization. Add SecurityAudit
REST controller and DTO. Add V35 Flyway migration for security_audits
table.
* feat(security): add scanner config to application profiles
Add scanner enabled flag to application-local.yml and
application-test.yml. Enable behavioral analyzer by default
in application.yml.
* feat(deploy): add skill-scanner to docker-compose and k8s manifests
Add skill-scanner service to docker-compose.yml with health check.
Add scanner k8s deployment, service, and configmap entries. Wire
scanner env vars into Makefile dev-all flow. Add verify-scanner.sh
script for post-deploy validation.
* docs(security): add scanner documentation suite
Add scanner docs: configuration guide, failure impact analysis,
monitoring guide, improvement recommendations, custom rules guide,
and skill-vetter rules conversion example. Update deployment docs
with scanner section. Add security-scanning overview and PRD.
* feat(security): add skill-vetter custom rule examples
Add example Regex and YARA rules derived from skill-vetter RED FLAGS
in scanner/examples/vetter-rules/. Includes 7 Regex rules
(signatures-append.yaml) and 3 YARA rules (skillhub_vetter.yara)
covering agent memory theft, IP-based exfiltration, and browser
data theft detection.
* feat(security): add scanner Docker build context
Add Dockerfile for cisco-ai-skill-scanner container and
.env.example with LLM configuration placeholders.
* fix(security): align Finding mapping with scanner API response schema
SkillScannerApiResponse.Finding used incorrect field names (message,
location.file, location.line, code_snippet) that did not match the
scanner's actual JSON output (description, file_path, line_number,
snippet), causing all four fields to deserialize as null.
Flatten Finding to match scanner API: remove nested Location, rename
fields to description/file_path/line_number/snippet. Add skill_name
and timestamp to SkillScannerApiResponse. Extend SecurityFinding with
remediation, analyzer, and metadata fields to capture LLM analyzer
output. Retain 8-arg compact constructor for backward compatibility.
* chore(security): add debug logging to scanner response mapping
Log raw scanner API response and mapped SecurityFinding fields
side-by-side to help verify data consistency between scanner
output and database records.
* feat(security): add multi-scanner support and soft delete for security audits
- Add ScannerType enum for type-safe scanner identification
- Update V35 migration to support multiple scanners and soft delete
- Remove CASCADE delete, use code-level soft delete (deleted_at)
- Add repository methods for querying latest audit by scanner type
- Update SecurityScanService to handle scanner type parameter
- Integrate soft delete in SkillHardDeleteService
- Update all tests to use ScannerType enum
This enables multiple scanner integrations (skill-scanner, future LLM/compliance scanners)
and preserves complete audit history through soft deletion.
* feat(security): add security audit UI to review detail and skill detail pages
Display security scan results on the review detail page (full audit
section with collapsible findings) and the skill detail sidebar (compact
summary with dialog for details). Handles empty/404 gracefully by
returning null, avoids loading shimmer flicker, and separates lifecycle
action buttons with a visual divider.
* docs(security): add security audit UI PRD
* docs(prd): add related documents section to file browser sidebar PRD
* feat(skill): add file tree builder utility
Implement buildFileTree function to convert flat SkillFile[] into hierarchical tree structure.
- Nodes sorted alphabetically by path
- Support for nested directories
- Each node includes depth, type, and path information
- Includes comprehensive test coverage
* feat(skill): add file type detection utilities
Implement utilities for file type detection and preview capability checks:
- isPreviewable: checks if file can be previewed based on extension and size
- canPreviewFile: provides detailed reason when file cannot be previewed
- getFileTypeLabel: returns human-readable file type labels
- getFileIcon: maps file types to appropriate Lucide icons
- Supports 1MB max file size limit
- Comprehensive test coverage for all functions
* feat(skill): add file tree node component
Implement recursive FileTreeNodeComponent with:
- Expand/collapse functionality for directories
- File and directory icons from Lucide React
- Hover effects showing file sizes
- Proper indentation based on depth
- Support for nested directory structures
* refactor(skill): upgrade file tree to hierarchical structure
Replace flat file list with tree structure:
- Use buildFileTree to convert flat files into hierarchy
- Integrate FileTreeNodeComponent for rendering
- Update onFileClick to accept FileTreeNode instead of SkillFile
- Add file count badge in header
- Root-level directories expanded by default
* feat(skill): add file preview dialog component
Implement FilePreviewDialog with:
- Markdown rendering via existing MarkdownRenderer
- Plain text/code display with monospace font
- Non-previewable file message with download button
- Copy-to-clipboard and download actions in header
- File path display in footer
- Loading and error states
* feat(i18n): add file tree and preview translations
Add Chinese and English translations for:
- File browser title
- Preview error messages (load failure, too large, binary, unsupported)
- Download file button
* feat(skill): integrate file preview into skill detail page
- Add useSkillFile hook for fetching arbitrary file content
- Add file preview state, click handler, and download handler
- Pass onFileClick to FileTree for opening preview dialog
- Add FilePreviewDialog to skill detail page
* feat(review): add file reading API endpoint
Add GET /api/v1/reviews/{id}/file?path=... for reading single files
from the review-bound skill version:
- ReviewController: new endpoint with path traversal validation
- GovernanceWorkflowAppService: route method for review file access
- ReviewSkillDetailAppService: authorization and delegation
- SkillQueryService: getFileContentByVersionId for direct version access
* feat(review): integrate file preview into review detail
- Add useReviewFile hook for fetching review file content
- Add file preview state and handlers to ReviewSkillDetailSection
- Pass reviewId prop for API calls
- Add FilePreviewDialog to review detail expanded section
- Update review-detail.tsx to pass taskId as reviewId
* fix(test): update test mocks for file preview hooks
- Add useSkillFile mock to skill-detail.test.tsx
- Add useQuery mock to @tanstack/react-query mock
- Add useReviewFile mock to review-skill-detail-section.test.tsx
- Fix SkillFile test fixtures to include all required properties
- Remove unused imports in test and component files
* fix(skill): use DomainBadRequestException for version not found
Replace DomainNotFoundException with DomainBadRequestException in
getFileContentByVersionId to match existing patterns and imports.
* feat(skill): add file tree sidebar with file sizes
- Move file tree to right sidebar for persistent visibility
- Display file sizes always visible (not just on hover)
- Keep Files tab showing the same tree structure
- Add scrollable container with max-height for long file lists
* fix(skill): improve file tree sidebar UI and UX
- Remove redundant title from file tree sidebar
- Adjust scrollbar placement with proper flex layout
- Sort folders first, then files (both alphabetically)
- Collapse all folders by default for cleaner initial view
- Increase dialog width to max-w-5xl and height to 90vh
- Remove duplicate close button (X icon)
- Add hover effects to action buttons
- Use flexbox for proper content scrolling
- Allow .pyc files in backend configuration
Fixes:
1. File tree sidebar now has cleaner layout without title
2. Scrollbar properly contained within card
3. Folders appear before files in tree
4. All folders collapsed by default
5. Preview dialog wider and more usable
6. Single close method (ESC or click outside)
7. Action buttons have visual feedback on hover
* fix(skill): refine file tree sidebar and preview dialog
1. Default all folders to collapsed state
2. File tree sidebar now uses Card+bare pattern matching SecurityAuditSummary
- Scrollbar inside the card content area instead of wrapping the tree
- Consistent padding and header layout with other sidebar cards
3. Preview dialog close button fixed:
- Hide DialogContent's built-in close button via [&>button]:hidden
- Add X button in the header row, same level as copy/download
- All action buttons share consistent hover animation (opacity transition)
4. Sort directories before files at every tree level
5. Update tests for new directory-first sort order
* feat(skill): collapsible file browser, wider dialog, button animations
1. File browser sidebar now supports expand/collapse toggle
- Click the header row to toggle visibility
- ChevronUp/Down icon indicates state
- Default expanded on page load
2. File preview dialog widened to 72rem (was 5xl/64rem)
- Uses w-[min(calc(100vw-2rem),72rem)] for responsive max
3. Header action buttons now have micro-interactions:
- Copy/Download: scale up on hover, scale down on click
- Close (X): rotates 90° on hover with destructive tint
- All buttons: opacity 60→100 transition on hover
* feat(skill): chevron animation, button effects, download path hint, review sidebar
1. File browser chevron: single ChevronDown with rotate-180 CSS transition
instead of swapping two icons — smooth 200ms rotation animation
2. Preview dialog action button animations:
- Copy icon: active:scale-125 feedback on click
- Download icon: hover:translate-y-0.5 subtle downward motion
- Close icon: hover:rotate-90 with destructive tint (unchanged)
3. Download tooltip now shows full file path:
"下载 src/prompts/system.md" instead of generic "下载文件"
- New i18n keys: filePreview.downloadHint, filePreview.copy, filePreview.close
4. Review detail page now has a sidebar file browser:
- Two-column layout (main content + lg:w-80 sidebar)
- File tree sourced from reviewSkillDetail.files (active review version)
- Collapsible with same chevron animation as skill detail page
- Shows active version badge below file tree
- File preview uses review file API (/reviews/{id}/file?path=...)
- Test mocks updated for new useReviewFile dependency
* fix(skill): download shows filename, copy rotates with success toast
1. Download button tooltip/text now shows actual filename:
"下载 README.md" instead of generic path
- i18n key changed from {{path}} to {{name}}
2. Copy button animation reworked to rotation:
- idle: hover:rotate-180 on the Copy icon
- click: animate-spin during clipboard write
- done: swap to green Check icon for 1.5s, then reset
3. Copy success toast notification:
- Shows "已复制到剪贴板" / "Copied to clipboard" via toast.success
- New i18n key: filePreview.copySuccess
* fix(skill): set download attribute so browser saves with original filename
The <a> element was missing the download attribute, causing the browser
to derive the filename from the URL path segment (/file?path=...) which
always resolved to "file". Now explicitly sets link.download = node.name
so the saved file uses the original name (e.g. README.md, config.json).
Applied to all three download handlers:
- skill-detail.tsx (skill version file)
- review-detail.tsx (review sidebar file)
- review-skill-detail-section.tsx (review expanded section file)
* feat(skill): add icons to all sidebar card headers
Add Lucide icons to each sidebar section for visual consistency
with the existing SecurityAuditSummary (Shield) and file browser (Folder):
- Terminal: Install command
- Clock: Pending review (amber-colored to match card theme)
- RefreshCw: Lifecycle management
- Tag: Label management (in SkillLabelPanel component)
- ArrowUpCircle: Promotion
- ShieldCheck: Governance
All icons use the same pattern: w-4 h-4 text-muted-foreground
placed inside a flex row with gap-2 before the section title.
* fix(skill): allow owners to preview all version statuses
When a skill has only a REJECTED version, the detail page crashed with
400 "版本未发布" because assertPreviewAccessible() only allowed PUBLISHED
and PENDING_REVIEW(owner). The frontend fallback to versions[0] would
pick the REJECTED version and fail all file/readme queries.
Backend changes:
- assertPreviewAccessible(): now allows owners and namespace admins to
preview ANY version status (DRAFT, REJECTED, YANKED, SCANNING, etc.)
via canManageRestrictedSkill() check. Also passes userNsRoles so
namespace admin role is respected, not just owner.
- resolveOwnerPendingPreview(): expanded from PENDING_REVIEW-only to
include all non-published, non-yanked statuses so headlineVersion
resolves correctly for owners with REJECTED/DRAFT/SCANNING versions.
Frontend changes:
- isOwnerPreviewResolution(): relaxed from checking PENDING_REVIEW
specifically to checking any non-PUBLISHED status, matching the
backend's broader projection.
Fixes: owner viewing skill with only REJECTED version → 400 error
Also fixes: DRAFT-only, SCANNING-only, SCAN_FAILED-only scenarios
* docs(requirements): add comprehensive requirements for file preview syntax highlighting
Generated complete documentation suite including:
- Core PRD (clarity score: 95/100)
- Constraints specification (8 business rules)
- Impact analysis (8 risks identified)
- Acceptance cases (25 test cases: 10 positive, 5 error, 7 boundary, 3 security)
- Test plan (100% coverage)
Mode: Quick (3 clarification rounds)
Complexity: Simple (3.5 days estimated)
* feat(skill): add syntax highlighting for code file preview
Implement syntax highlighting for code files in the file preview dialog:
- Add CodeRenderer component using lowlight (highlight.js wrapper via rehype-highlight)
- Add getLanguageForHighlight() to map file extensions to highlight.js languages
- Support 20+ languages: Python, JS/TS, Java, Go, Rust, C/C++, Ruby, PHP, Shell, JSON, YAML, XML, etc.
- Apply 500KB threshold for syntax highlighting (larger files show plain text)
- Maintain visual consistency with Markdown code blocks
- Auto-adapt to light/dark theme using existing CSS variables
- Add lowlight as direct dependency to enable programmatic syntax highlighting
Implements solution 1 from PRD (reuse rehype-highlight infrastructure).
* fix(security): resolve security audit FK constraint and stale query issues
Backend:
- Add hardDeleteByVersionId to physically remove audit records before
deleting skill versions (fixes FK constraint violation on hard delete)
- Add softDeleteByVersionId calls in SkillGovernanceService and
SkillPublishService version deletion paths
- Add V36 migration to convert security_audit TIMESTAMP columns to
TIMESTAMPTZ matching project convention
- Update tests to verify audit cleanup in all deletion flows
Frontend:
- Prevent stale API calls after skill deletion by gating all query
hooks with a skillDeleted flag that disables them immediately
- Move cache cleanup from useDeleteSkill onSuccess to the handler
to avoid refetching while the component is still mounted
* fix(skill): handle SCANNING and SCAN_FAILED statuses across all pages
Backend:
- Include SCANNING/SCAN_FAILED in version listing filters so owners
can see versions being scanned
- Add explicit lifecycle sort priority for scanner statuses
- Allow deletion of SCAN_FAILED versions alongside DRAFT and REJECTED
Frontend:
- Add i18n labels and CSS classes for SCANNING/SCAN_FAILED in my-skills
- Add localized version status labels in skill-detail version list
- Hide archive button when no published version exists (my-skills and
skill-detail)
- Allow deleting SCAN_FAILED versions from skill-detail
* gitignore
|
||
|---|---|---|
| .. | ||
| 02-constraints.md | ||
| 03-impact-analysis.md | ||
| 04-acceptance-cases.md | ||
| 05-test-plan.md | ||
| README.md | ||
文件预览语法高亮 - 需求文档索引
文档概览
本目录包含"文件预览语法高亮"功能的完整需求文档套件。
功能简介：为技能包文件预览功能添加语法高亮支持，提升代码可读性。
创建日期：2026-03-22 文档版本：1.0 清晰度评分：95/100 预计工作量：3.5 天
核心文档
📋 产品需求文档 (PRD)
类型：核心 PRD 生成方式：requirements-clarity 技能 内容：
- 需求描述（背景、功能概述、详细需求）
- 设计决策（技术方案、约束条件、风险评估）
- 验收标准（功能验收、质量标准、用户验收）
- 执行阶段（准备、开发、测试、部署）
- 后端优化方案（后续实施）
补充文档
🔒 约束规范
描述：定义功能边界、业务规则、技术约束和依赖关系 关键内容：
- 功能边界：包含/排除/未来可能性
- 业务规则：8 条规则（BR-001 ~ BR-008）
- 技术约束：性能、兼容性、安全、可扩展性
- 依赖关系：上游依赖、下游影响、外部依赖
- 非功能性需求：可用性、可维护性、可测试性、可观测性
📊 影响分析
描述：代码影响范围、API 变更、数据库变更、风险评估 关键内容：
- 代码影响矩阵：8 个模块，主要影响前端
- API 影响：无新增/修改端点，复用现有 API
- 数据库影响：无模式变更
- 风险评估：8 个风险（R-001 ~ R-008）
- 回滚计划：触发条件、回滚步骤、影响范围
✅ 验收用例
描述：详细的测试场景，包括正向路径、错误路径、边界条件和安全测试 关键内容：
- 正向路径：10 个用例（AC-P-001 ~ AC-P-010）
- 错误路径：5 个用例（AC-E-001 ~ AC-E-005）
- 边界条件：7 个用例（AC-B-001 ~ AC-B-007）
- 安全测试：3 个用例（AC-S-001 ~ AC-S-003）
- 总计：25 个验收用例
🧪 测试计划
描述：单元测试、集成测试、性能测试的详细计划和覆盖率矩阵 关键内容：
- 单元测试：2 个测试类，16 个测试方法
- 集成测试：2 个测试类，11 个测试方法
- 性能测试：4 个场景（渲染时间、内存、包体积）
- 浏览器兼容性测试：4 个浏览器
- 安全测试：3 个场景（XSS 防护）
- 覆盖率：100%（25/25 验收用例）
业务规则速查
| 规则 ID | 描述 | 优先级 |
|---|---|---|
| BR-001 | 文件大小 > 500KB → 纯文本（无高亮） | Must |
| BR-002 | 文件大小 > 1MB → 仅下载 | Must |
| BR-003 | 无法识别语言 → 纯文本（不报错） | Must |
| BR-004 | 渲染失败 → 降级到纯文本 | Must |
| BR-005 | 内存不足 → 降级到纯文本 | Should |
| BR-006 | 主题跟随系统 | Must |
| BR-007 | 样式与 Markdown 一致 | Must |
| BR-008 | 复制功能保留 | Must |
风险速查
| 风险 ID | 描述 | 概率 | 影响 | 缓解措施 |
|---|---|---|---|---|
| R-001 | 大文件卡顿 | 中 | 高 | 500KB 阈值 |
| R-005 | 渲染失败崩溃 | 低 | 高 | Error Boundary |
| R-007 | 后端性能下降 | 中 | 中 | 后续缓存+限流 |
| R-008 | XSS 安全风险 | 低 | 高 | 确保转义 |
实施路线图
阶段 1：准备工作（0.5 天）
- 调研 highlight.js 按需导入方案
- 验证样式复用性
- 设计 CodeRenderer 组件 API
阶段 2：核心开发（1.5 天）
- 创建 CodeRenderer 组件
- 扩展 file-type-utils.ts
- 修改 file-preview-dialog.tsx
- 主题适配
阶段 3：集成与测试（1 天）
- 单元测试（16 个测试方法）
- 集成测试（11 个测试方法）
- 性能测试（4 个场景）
- 浏览器兼容性测试
阶段 4：部署与监控（0.5 天）
- 代码审查（安全、性能）
- 部署到生产环境
- 监控指标配置
- 文档更新
后续优化（不在本次范围）
P1（下个迭代）
- 后端 Redis 缓存（缓存命中率 > 60%）
P2（后续优化）
- 限流策略（认证 60 次/分钟，匿名 20 次/分钟）
- 监控指标（响应时间、错误率、缓存命中率）
文档使用指南
开发阶段
- 开始开发前：阅读 PRD 和约束规范，理解功能边界和业务规则
- 编写代码时：参考影响分析，了解需要修改的文件
- 编写测试时：参考验收用例和测试计划，确保覆盖所有场景
- 代码审查时：检查是否满足所有业务规则（BR-XXX）
测试阶段
- 单元测试：按照测试计划执行，确保覆盖率 > 80%
- 集成测试：验证所有验收用例（AC-XXX）
- 性能测试：确保满足性能约束（500KB < 500ms）
- 安全测试：验证 XSS 防护（AC-S-XXX）
部署阶段
- 部署前：检查回滚计划是否就绪
- 部署后：监控错误率和性能指标
- 问题排查：参考风险评估（R-XXX）定位问题
变更日志
| 日期 | 文档 | 变更 | 原因 |
|---|---|---|---|
| 2026-03-22 | 所有文档 | 初始版本创建 | 需求澄清完成（3 轮交互） |
联系方式
如有疑问，请联系：
- 产品负责人：待定
- 技术负责人：待定
- 测试负责人：待定