feat: add security guides and sanitize external links across repo

New articles:
- the-security-guide.md: "The Shorthand Guide to Securing Your Agent" (595 lines)
  Attack vectors, sandboxing, sanitization, OWASP Top 10, observability
- the-openclaw-guide.md: "The Hidden Danger of OpenClaw" (470 lines)
  Security analysis of OpenClaw, MiniClaw thesis, industry evidence

External link sanitization (22 files across EN, zh-CN, zh-TW, ja-JP, .cursor):
- Removed third-party GitHub links from skills and guides
- Replaced with inline descriptions to prevent transitive prompt injection
- Kept official org links (Anthropic, Google, Supabase, Mixedbread)

docs/zh-CN/skills/continuous-learning-v2/SKILL.md

@@ -282,7 +282,7 @@ v2 与 v1 完全兼容：
 ## 相关链接
 
 * [技能创建器](https://skill-creator.app) - 从仓库历史生成本能
-* [Homunculus](https://github.com/humanplane/homunculus) - v2 架构的灵感来源
+* Homunculus - 启发 v2 架构的社区项目（原子观察、置信度评分、本能演化管线）
 * [长文指南](https://x.com/affaanmustafa/status/2014040193557471352) - 持续学习部分
 
 ***

docs/zh-CN/skills/continuous-learning/SKILL.md

@@ -83,7 +83,7 @@ description: 自动从Claude Code会话中提取可重用模式，并将其保
 
 ## 对比说明（研究：2025年1月）
 
-### 与 Homunculus (github.com/humanplane/homunculus) 对比
+### 与 Homunculus 对比
 
 Homunculus v2 采用了更复杂的方法：
 

docs/zh-CN/skills/nutrient-document-processing/SKILL.md

@@ -161,5 +161,4 @@ curl -X POST https://api.nutrient.io/build \
 
 * [API 演练场](https://dashboard.nutrient.io/processor-api/playground/)
 * [完整 API 文档](https://www.nutrient.io/guides/dws-processor/)
-* [代理技能仓库](https://github.com/PSPDFKit-labs/nutrient-agent-skill)
 * [npm MCP 服务器](https://www.npmjs.com/package/@nutrient-sdk/dws-mcp-server)