claude-dash-admin

Compare original and translation side by side

🇺🇸

Original

English

🇨🇳

Translation

Chinese

Claude-Dash Administration Framework

Claude-Dash 管理框架

When This Activates

触发场景

This skill activates for system maintenance:

Health checks and diagnostics
Storage cleanup and optimization
Index management
Service troubleshooting

该技能在以下系统维护场景下激活：

健康检查与诊断
存储清理与优化
索引管理
服务故障排查

System Components

系统组件

Core Services

核心服务

Service	Purpose	Check Command
Watcher	File monitoring, index updates	`ps aux \| grep watcher`
Gateway	MCP routing, caching	`ps aux \| grep gateway`
Ollama	Local LLM	`ollama list`
Dashboard	Visualization	Port 3333

服务	用途	检查命令
Watcher	文件监控、索引更新	`ps aux \| grep watcher`
Gateway	MCP路由、缓存	`ps aux \| grep gateway`
Ollama	本地LLM	`ollama list`
Dashboard	可视化展示	端口3333

Storage Locations

存储位置

~/.claude-dash/
├── projects/       # Project memory (indexes, summaries)
├── sessions/       # Session data, transcripts
├── learning/       # Corrections, preferences, calibration
├── mlx-tools/      # Local AI tools (Ollama-based)
├── skills/         # Skill definitions
├── logs/           # Service logs
├── indexes/        # HNSW search indexes
└── config.json     # Main configuration

~/.claude-dash/
├── projects/       # 项目内存（索引、摘要）
├── sessions/       # 会话数据、对话记录
├── learning/       # 修正记录、偏好设置、校准数据
├── mlx-tools/      # 本地AI工具（基于Ollama）
├── skills/         # 技能定义
├── logs/           # 服务日志
├── indexes/        # HNSW搜索索引
└── config.json     # 主配置文件

Health Checks

健康检查

Quick Health Check

快速健康检查

memory_health action=status

memory_health action=status

Full Diagnostic

完整诊断

bash

undefined

bash

undefined

Check services

检查服务状态

ps aux | grep -E "(watcher|gateway|ollama)"

Check storage

检查存储占用

du -sh ~/.claude-dash/*/

Check recent errors

查看近期错误日志

tail -20 ~/.claude-dash/logs/watcher-error.log

Check gateway metrics

查看网关指标

gateway_metrics format=summary

undefined

gateway_metrics format=summary

undefined

Common Maintenance Tasks

常见维护任务

1. Clean Up Storage

1. 存储清理

bash

undefined

bash

undefined

Run session archival

执行会话归档

python3 ~/.claude-dash/scripts/archive-sessions.py

Run log rotation

执行日志轮转

~/.claude-dash/scripts/log-rotation.sh

undefined

~/.claude-dash/scripts/log-rotation.sh

undefined

2. Refresh Indexes

2. 刷新索引

bash

undefined

bash

undefined

Trigger HNSW rebuild

触发HNSW索引重建

hnsw_status action=rebuild project=gyst

Run freshness check

执行新鲜度检查

workers_run worker=freshness project=gyst

undefined

workers_run worker=freshness project=gyst

undefined

3. Consolidate Learning

3. 整合学习数据

bash

undefined

bash

undefined

Run observation consolidation

执行观测数据整合

workers_run worker=consolidate

undefined

workers_run worker=consolidate

undefined

4. Restart Services

4. 重启服务

bash

undefined

bash

undefined

Restart watcher

重启Watcher

~/.claude-dash/watcher/start-watcher.sh restart

Restart gateway (via Claude settings or restart Claude)

重启网关（通过Claude设置或重启Claude）

undefined

undefined

Troubleshooting

故障排查

"Memory search returns nothing"

“内存搜索无结果”

Check if project is registered in config.json
Verify index files exist:
```
ls ~/.claude-dash/projects/{project}/
```
Check watcher is running
Trigger reindex if needed

检查项目是否在config.json中注册
验证索引文件是否存在：
```
ls ~/.claude-dash/projects/{project}/
```
检查Watcher是否在运行
必要时触发重新索引

"Hook is slow"

“Hook响应缓慢”

Check hook timeout in settings
Review inject-context.sh performance
Consolidate to single Python process (inject_all_context.py)

检查设置中的Hook超时时间
评估inject-context.sh的性能
合并为单个Python进程（使用inject_all_context.py）

"Watcher crashes"

“Watcher崩溃”

Check watcher-error.log
Verify project paths exist
Check for permission issues

查看watcher-error.log日志
验证项目路径是否存在
检查权限问题

"Ollama not responding"

“Ollama无响应”

Check:
```
ollama list
```
Restart:
```
ollama serve
```
Verify model exists:
```
ollama pull gemma3:4b-it-qat
```

检查状态：
```
ollama list
```
重启服务：
```
ollama serve
```
验证模型是否存在：
```
ollama pull gemma3:4b-it-qat
```

Configuration

配置

Main Config (config.json)

主配置文件（config.json）

json

{
  "projects": [...],
  "watcher": {
    "enabled": true,
    "ignorePatterns": [...],
    "scanIntervalMs": 5000
  }
}

json

{
  "projects": [...],
  "watcher": {
    "enabled": true,
    "ignorePatterns": [...],
    "scanIntervalMs": 5000
  }
}

Adding a New Project

添加新项目

json

{
  "id": "project-id",
  "displayName": "Project Name",
  "path": "/full/path/to/project",
  "memoryPath": "projects/project-id"
}

json

{
  "id": "project-id",
  "displayName": "Project Name",
  "path": "/full/path/to/project",
  "memoryPath": "projects/project-id"
}

Performance Metrics

性能指标

Gateway Efficiency

网关效率

gateway_metrics format=detailed

Shows:

Cache hit rate (target: >50%)
Local routing rate (higher = cheaper)
Token savings

gateway_metrics format=detailed

展示内容：

缓存命中率（目标：>50%）
本地路由率（数值越高成本越低）
Token节省量

Learning Status

学习状态

learning_status component=all

Shows:

Preferences learned
Corrections recorded
Confidence calibration

learning_status component=all

展示内容：

已学习的偏好设置
已记录的修正内容
置信度校准数据

Scheduled Maintenance

定期维护

These run automatically via cron:

3 AM: Log rotation
Background: Watcher auto-updates

Manual periodic tasks:

Weekly: Session archival
Monthly: Review and clean old projects
As needed: HNSW index rebuild

以下任务通过cron自动执行：

凌晨3点：日志轮转
后台运行：Watcher自动更新

手动定期任务：

每周：会话归档
每月：审核并清理旧项目
按需：HNSW索引重建

Emergency Recovery

紧急恢复

Reset Project Memory

重置项目内存

bash

rm -rf ~/.claude-dash/projects/{project}/*

bash

rm -rf ~/.claude-dash/projects/{project}/*

Watcher will rebuild on next file change

Watcher会在下次文件变更时重建索引

undefined

undefined

Reset All Indexes

重置所有索引

bash

hnsw_status action=rebuild-all

bash

hnsw_status action=rebuild-all

Clear Corrupted State

清除损坏状态

bash

undefined

bash

undefined

Backup first

先备份数据

cp -r ~/.claude-dash ~/.claude-dash.backup

Then clear specific problematic data

然后清除特定损坏数据

rm ~/.claude-dash/sessions/observations.json rm ~/.claude-dash/learning/corrections.json

undefined

rm ~/.claude-dash/sessions/observations.json rm ~/.claude-dash/learning/corrections.json

undefined