系统配置仪表板_diverga-育师

以下为本文档的中文说明

该技能用于Diverga仪表板——提供实时的配置状态和功能概览。Diverga是一个系统管理平台，此技能帮助用户快速查看系统的运行配置、功能开关状态和关键性能指标。它提供了一目了然的可视化仪表板，整合分散的系统信息于统一的视图之中。适用于系统管理员和运维工程师需要实时了解系统运行状态的场景。该技能通过集中化的信息展示，减少了在多个管理界面之间切换的时间成本，帮助运维团队快速发现配置异常和功能问题，提升系统管理的效率和响应速度。该技能提供了详细的操作指南和最佳实践，帮助用户快速上手并深入掌握。通过系统的功能模块划分和丰富的应用场景说明，用户可以在实际项目中有效运用该技能提升工作效率。该技能注重实用性和可操作性，涵盖从基础配置到高级功能的完整知识体系，满足不同层次用户的学习需求。持续更新和优化的内容确保用户始终能够接触到最新的技术发展和行业实践。通过此技能的学习和应用，用户可以减少摸索时间，快速获得可用的解决方案，将精力集中在核心业务逻辑和创新工作上，从而在技术快速迭代的环境中保持竞争力。该技能的模块化设计使其易于扩展和定制，用户可以根据自身需求灵活调整应用方式，实现最大化的价值产出。该技能整合了常见的设计模式和最佳实践，提供了清晰的学习路径和参考资料，帮助用户在短时间内建立起完整的知识框架，并有能力在实际项目中灵活运用所学内容解决问题。

Diverga Dashboard

When the user runs/diverga(the base command), display a live configuration dashboard.

Instructions

Display the following dashboard by reading live configuration and environment state. Use plain text output (no tool calls needed for display — just output the text directly).

Step 1: Display ASCII Logo

Output this exact logo block:

╔═══════════════════════════════════════════════════════════════════════════════╗ ║ ║ ║ ██████╗ ██╗██╗ ██╗███████╗██████╗ ██████╗ █████╗ ║ ║ ██╔══██╗██║██║ ██║██╔════╝██╔══██╗██╔════╝ ██╔══██╗ ║ ║ ██║ ██║██║██║ ██║█████╗ ██████╔╝██║ ███╗███████║ ║ ║ ██║ ██║██║╚██╗ ██╔╝██╔══╝ ██╔══██╗██║ ██║██╔══██║ ║ ║ ██████╔╝██║ ╚████╔╝ ███████╗██║ ██║╚██████╔╝██║ ██║ ║ ║ ╚═════╝ ╚═╝ ╚═══╝ ╚══════╝╚═╝ ╚═╝ ╚═════╝ ╚═╝ ╚═╝ ║ ║ ║ ║ * Diverge from the Modal · Discover the Exceptional ║ ║ ║ ╚═══════════════════════════════════════════════════════════════════════════════╝

Step 2: Read Configuration

Readconfig/diverga-config.json(relative to the plugin directory at~/.claude/plugins/diverga/config/diverga-config.json). Extract:

human_checkpoints.levelor infer fromhuman_checkpoints.enabled+human_checkpoints.required
hud.presetif present
language
version

If config file doesn’t exist, show “Not configured — run /diverga:setup”

Step 3: Check Project Status

Check if.research/project-state.yamlexists in the current working directory.

If exists: read project name, current stage, and checkpoint status
If not: show “No active project”

Step 4: Check API/MCP Status

Check these environment variables and report status:

SEMANTIC_SCHOLAR_API_KEY— for paper retrieval
OPENALEX_EMAIL— for OpenAlex polite pool
SCOPUS_API_KEY— for Scopus access
WOS_API_KEY— for Web of Science access
GROQ_API_KEY— for LLM screening (recommended)
ANTHROPIC_API_KEY— for Claude API screening
OPENAI_API_KEY— for OpenAI screening
GEMINI_API_KEY— for Gemini visualization

For each: show “configured” if set, “not set” if unset, with setup hint.

Step 5: Display Dashboard

Combine all information into this format:

SYSTEM STATUS Version: 11.0.0 Agents: 24 agents across 9 categories (A-G, I, X) Status: Ready CONFIGURATION Checkpoint: [level from config or "Full (default)"] HUD Preset: [preset from config or "research (default)"] VS Method: Enabled PROJECT Active: [project name or "No active project"] Stage: [stage or "-"] Memory: [healthy/no project] API STATUS (Paper Retrieval) Semantic Scholar: [configured/not set -> export SEMANTIC_SCHOLAR_API_KEY=...] OpenAlex: [configured (email: ...)/not set -> export OPENALEX_EM AIL=your@email.com] Scopus: [configured/not set -> export SCOPUS_API_KEY=...] Web of Science: [configured/not set -> export WOS_API_KEY=...] LLM PROVIDERS (Batch Screening via I2) Groq: [configured/not set -> export GROQ_API_KEY=...] llama-3.3-70b — $0.01/100 papers (recommended) Anthropic API: [configured/not set -> export ANTHROPIC_API_KEY=...] claude-haiku-4-5 — $0.15/100 papers OpenAI: [configured/not set -> export OPENAI_API_KEY=...] gpt-4o-mini — ~$0.02/100 papers MCP INTEGRATIONS Zotero: [detected/not available -> /diverga:setup for guidance] Claude Code Browser: [detected/not available] Gemini: [configured/not set -> export GEMINI_API_KEY=...] QUICK ACTIONS /diverga:setup Initial configuration /diverga:doctor System diagnostics /diverga:help All 24 agents & commands /diverga:memory Project context & status /diverga:hud HUD display settings "Start a systematic review on [topic]" -> Begin research

Step 6: Offer MCP Setup Guidance (if requested)

If the user asks about Zotero MCP setup, provide:

Zotero MCP enables: - Auto-import retrieved papers into Zotero library - Collection management during screening (Included/Excluded) - Bibliography generation in any citation style - Citation insertion during manuscript writing Setup: 1. Install Zotero desktop app (zotero.org) 2. Install Zotero MCP server 3. Add to Claude Code MCP config 4. Restart Claude Code

If the user asks about Claude Code Browser:

Claude Code Browser enables: - Automated search queries on institutional databases (WoS, Scopus, ERIC) - Guided export from university-licensed portals - Semi-automated paper collection from any web-based database Status: Available if Chrome extension is installed. No additional setup needed.

Implementation Notes

Use Bash tool to check environment variables:echo $VARIABLE_NAME
Use Read tool to check config file existence and contents
Use Glob tool to check for.research/project-state.yaml
Display everything as formatted text output
The dashboard is READ-ONLY — it doesn’t modify any files
If any check fails, show graceful fallback (e.g., “Unable to determine” instead of errors)