query-dot-ai
Query sibling dot-ai projects to verify features are USABLE (not just defined). IMPORTANT: When calling this skill, explain HOW you plan to use the feature (e.g., 'I need to call X via REST API from the UI' or 'I need to import Y function'). This helps verify the full chain from definition to exposure.
安装与使用
请根据来源仓库和你的 AI 工具说明安装此技能。先阅读技能文件和依赖要求,再添加到工作流。
查看来源 ↗使用指南技能说明
Query dot-ai Projects
Explore the dot-ai ecosystem codebases to find the requested information.
Project Locations
Sibling projects are located in the parent directory of the current working directory (../):
- dot-ai - Main MCP server (API endpoints, tools, handlers)
- dot-ai-ui - Web UI for visualizations and dashboard
- dot-ai-controller - Kubernetes controller
- dot-ai-stack - Stack deployment configs
- dot-ai-website - Documentation website
Default to dot-ai (MCP server) if the target project is unclear.
Important: Do NOT use this skill to query the project you're currently working in. Use local tools (Read, Grep, Glob) instead.
Excluded
dot-ai-infra - Production infrastructure. Only query if user explicitly requests it.
Verification Mindset
Don't just find that something EXISTS - prove it's USABLE.
- Finding a type/interface is NOT enough
- Finding internal code is NOT enough
- You must trace from definition → implementation → exposure
When asked "does X exist?", answer:
- "Yes, and here's how to use it: [concrete usage]" OR
- "It exists internally but is NOT exposed for external use"
Go deep, not wide. Follow the code path until you can prove how the caller would actually use the feature.
使用示例
查看来源仓库中的使用示例与配置说明。
查看来源 ↗来源历史
- 收录时间
- 2026-01-20
- 索引更新时间
- 2026-02-04
- Branch
- main
收录与同步日期属于本站记录。源码修改历史请查看来源仓库。
GitHub ↗chat Comments (0)
Sign in to join the discussion and leave a comment.
No comments yet. Be the first to share your thoughts!