Add enhanced onboarding examples and Xilope production guide

**Onboarding Enhancements**:
- Added explicit server discovery examples
- Included Claude Code server usage patterns
- Added read_file tool call examples with proper namespacing

**Xilope Production Guide**:
- Complete production environment documentation
- Memory storage integration details (/mnt/data/claude/claude/.mcp-memory/)
- cmem wrapper integration (~/bin/cmem → /usr/local/bin/cmem)
- Bidirectional sync workflows and best practices

**Server Updates**:
- Enhanced onboarding server to serve predefined markdown files
- Priority: predefined files → default.md → fallback instructions

This provides comprehensive guidance for both general MCP Browser usage
and Xilope's specific production environment with cmem integration.

🤖 Generated with [Claude Code](https://claude.ai/code)

Co-Authored-By: Claude <noreply@anthropic.com>
This commit is contained in:
Claude4Ξlope 2025-06-28 12:52:48 +02:00
parent 63c26975c4
commit c55ffdb59b
3 changed files with 277 additions and 9 deletions

View File

@ -46,6 +46,17 @@ mcp_call(method="tools/call", params={"name": "brave_web_search", "arguments": {
## Discovery Examples
### Listing Available Servers
```python
# First, discover what MCP servers are available
mcp_discover(jsonpath="$.servers[*].name")
# Returns: ["builtin", "claude-code", "filesystem", etc.]
# Get detailed info about servers
mcp_discover(jsonpath="$.servers[*]")
```
### Basic Tool Discovery
```python
# Discover all available tools (built-in + external servers)
mcp_discover(jsonpath="$.tools[*].name")
@ -53,13 +64,20 @@ mcp_discover(jsonpath="$.tools[*].name")
# Get tools from specific server
mcp_discover(jsonpath="$.servers.brave-search.tools[*].name")
# Get all configured servers
mcp_discover(jsonpath="$.servers[*].name")
# Get tool details
mcp_discover(jsonpath="$.tools[?(@.name=='brave_web_search')]")
```
### Server-Specific Tool Discovery
```python
# Example: Get all Claude Code tools
mcp_discover(jsonpath="$.servers['claude-code'].tools[*].name")
# Returns: ["read_file", "write_file", "list_directory", etc.]
# Get details of a specific tool from claude-code server
mcp_discover(jsonpath="$.servers['claude-code'].tools[?(@.name=='read_file')]")
```
## Using External Server Tools
Once discovered, call any tool through `mcp_call`:
@ -82,6 +100,29 @@ mcp_call(
"arguments": {"query": "mcp-browser"}
}
)
# Example: Using Claude Code server to read a file
mcp_call(
method="tools/call",
params={
"name": "claude-code::read_file",
"arguments": {
"path": "/path/to/file.py",
"start_line": 1,
"end_line": 50
}
}
)
# Alternative syntax for server-namespaced tools
mcp_call(
method="tools/call",
params={
"name": "read_file",
"arguments": {"path": "/path/to/file.py"},
"server": "claude-code"
}
)
```
## Runtime Configuration

View File

@ -125,17 +125,21 @@ class OnboardingServer(BaseMCPServer):
content = self._format_onboarding(identity, data)
return self.content_text(content)
else:
# Try to load predefined markdown files first
predefined_file = Path(__file__).parent / f"{identity}.md"
if predefined_file.exists():
with open(predefined_file) as f:
predefined_content = f.read()
return self.content_text(predefined_content)
# Try to load default onboarding
default_file = self.onboarding_dir / "default.md"
default_file = Path(__file__).parent / "default.md"
if default_file.exists():
with open(default_file) as f:
default_content = f.read()
return self.content_text(
f"# Onboarding for {identity}\n\n"
f"No specific onboarding found. Using default:\n\n"
f"{default_content}"
)
return self.content_text(default_content)
else:
return self.content_text(
f"# Onboarding for {identity}\n\n"

View File

@ -0,0 +1,223 @@
# Xilope Production Environment - MCP Browser Guide
Welcome to Xilope's production MCP Browser setup! This environment provides integrated access to both MCP tools and the Claude Memory (cmem) system.
## Production Architecture
### Memory Storage Integration
- **Local Storage**: `/mnt/data/claude/claude/.mcp-memory/`
- **cmem Integration**: Bidirectional sync via `/mnt/data/claude/claude/bin/cmem`
- **Identity-Based**: Each project gets separate memory space
- **Persistent**: Memory survives across AI assistant sessions
### Built-in Servers Available
1. **Memory Server** (`builtin:memory::`):
- `task_add`, `task_list`, `task_update` - Task management with cmem sync
- `decision_add` - Decision tracking with reasoning
- `pattern_add`, `pattern_resolve` - Learning pattern management
- `knowledge_add`, `knowledge_get` - Fact storage and retrieval
- `project_switch` - Switch between project contexts
- `memory_summary` - Get overview of stored information
2. **Screen Server** (`builtin:screen::`):
- `create_session`, `execute`, `peek` - GNU screen management
- `list_sessions`, `kill_session` - Session lifecycle
- `enable_multiuser`, `attach_multiuser`, `add_user` - Collaboration
3. **Pattern Server** (`builtin:patterns::`):
- `add_pattern`, `list_patterns` - Auto-response pattern management
- `test_pattern`, `execute_pattern` - Pattern execution
4. **Onboarding Server** (`builtin:onboarding::`):
- `onboarding` - Identity-aware instructions
- `onboarding_list`, `onboarding_delete`, `onboarding_export` - Management
## Production Workflows
### Starting a Session
```python
# Check what servers are available
mcp_discover(jsonpath="$.servers[*].name")
# List all memory tools for task management
mcp_discover(jsonpath="$.tools[?(@.name =~ /memory|task/)]")
# Check current project context
mcp_call(
method="tools/call",
params={
"name": "builtin:memory::memory_summary",
"arguments": {}
}
)
```
### Task Management with cmem Sync
```python
# Add a new task (automatically syncs to cmem)
mcp_call(
method="tools/call",
params={
"name": "builtin:memory::task_add",
"arguments": {
"content": "Implement feature X",
"priority": "high",
"assignee": "next-ai"
}
}
)
# List active tasks
mcp_call(
method="tools/call",
params={
"name": "builtin:memory::task_list",
"arguments": {"status": "pending"}
}
)
# Update task status (syncs completion to cmem)
mcp_call(
method="tools/call",
params={
"name": "builtin:memory::task_update",
"arguments": {
"task_id": "abc123",
"status": "completed"
}
}
)
```
### Decision and Pattern Management
```python
# Record important decisions (synced to cmem)
mcp_call(
method="tools/call",
params={
"name": "builtin:memory::decision_add",
"arguments": {
"choice": "Use Docker for deployment",
"reasoning": "Simplifies environment management",
"alternatives": ["Native install", "VM deployment"]
}
}
)
# Add learning patterns (synced to cmem)
mcp_call(
method="tools/call",
params={
"name": "builtin:memory::pattern_add",
"arguments": {
"pattern": "Always test before commit",
"description": "Run full test suite before any git commit",
"priority": "high"
}
}
)
```
### Screen Session Management
```python
# Create a development session
mcp_call(
method="tools/call",
params={
"name": "builtin:screen::create_session",
"arguments": {
"session_name": "development",
"working_directory": "/mnt/data/claude/claude"
}
}
)
# Execute commands in session
mcp_call(
method="tools/call",
params={
"name": "builtin:screen::execute",
"arguments": {
"session_name": "development",
"command": "git status"
}
}
)
# Peek at session output
mcp_call(
method="tools/call",
params={
"name": "builtin:screen::peek",
"arguments": {"session_name": "development"}
}
)
```
### Project Context Switching
```python
# Switch to different project memory space
mcp_call(
method="tools/call",
params={
"name": "builtin:memory::project_switch",
"arguments": {"project_name": "mcp-browser"}
}
)
# Use onboarding for project-specific instructions
mcp_call(
method="tools/call",
params={
"name": "builtin:onboarding::onboarding",
"arguments": {
"identity": "mcp-browser",
"instructions": "Focus on context optimization and AI-first development"
}
}
)
```
## cmem Integration Details
### Memory Storage Structure
```
/mnt/data/claude/claude/.mcp-memory/
├── default/ # Default project space
│ ├── tasks.json # Task storage
│ ├── decisions.json # Decision history
│ ├── patterns.json # Learning patterns
│ └── knowledge.json # Knowledge base
├── mcp-browser/ # Project-specific space
└── [other-projects]/ # Additional projects
```
### cmem Wrapper
- **Location**: `/mnt/data/claude/claude/bin/cmem`
- **Function**: Wraps `/usr/local/bin/cmem` with proper directory context
- **Integration**: Automatic bidirectional sync with MCP memory server
- **Commands**: `cmem handoff`, `cmem task add`, `cmem pattern add`, etc.
### Sync Behavior
- **Automatic**: All memory operations sync to cmem in background
- **Graceful**: If cmem unavailable, operations continue locally
- **Identity-Aware**: Each project gets separate cmem context
- **Bidirectional**: Changes in either system propagate to the other
## Production Best Practices
1. **Always check memory summary** at start of session
2. **Use task management** to track work across AI sessions
3. **Record decisions** with reasoning for future reference
4. **Create patterns** to capture effective approaches
5. **Switch project contexts** when working on different codebases
6. **Use screen sessions** for persistent development environments
## Error Handling
If cmem sync fails:
- Operations continue with local storage
- Sync retries automatically when cmem becomes available
- No data loss occurs during temporary cmem unavailability
This production setup ensures seamless AI assistant transitions while maintaining full project context and memory across sessions.