@@ -47,219 +47,10 @@ language model and flexible integration capabilities through the Model Context P
4747allows us to create custom workflows and integrate with our existing development tools while
4848maintaining control over data privacy and security.
4949
50- ## Installing Claude Code and Claude Desktop
50+ :::tip Setting up AI tooling
5151
52- ### Claude Code
52+ To set up AI tooling in your development environment, see the
53+ [ AI Tools] ( ./../getting-started/tools/index.md#ai-tools ) instructions of our Getting Started
54+ section.
5355
54- Claude Code is Anthropic's official CLI tool that brings Claude's capabilities directly to your
55- terminal. It's ideal for developers who prefer command-line interfaces and want to integrate AI
56- assistance into their terminal-based workflows.
57-
58- #### Installation
59-
60- 1 . [ Node.js] ( https://nodejs.org/ ) v18 or higher is available
61- 2 . Install via NPM ` npm install -g @anthropic-ai/claude-code ` or Homebrew
62- ` brew install --cask claude-code `
63- 3 . Configure your API key:
64-
65- ``` bash
66- claude configure
67- ```
68-
69- Walk through the process to sign into the Anthropic Console via SSO and authenticate your local
70- client.
71-
72- #### Basic usage
73-
74- ``` bash
75- # Start an interactive session
76- claude
77-
78- # Ask a question
79- claude " How do I add a feature flag around my changes?"
80- ```
81-
82- ### Claude Desktop
83-
84- Claude Desktop provides a graphical interface for interacting with Claude, ideal for developers who
85- prefer a dedicated application with rich formatting and file management capabilities.
86-
87- #### Installation
88-
89- Install via [ claude.ai/download] ( https://claude.ai/download ) or Homebrew ` brew install claude `
90-
91- - Launch Claude Desktop
92- - Sign in with your Anthropic account via SSO
93- - Configure your workspace preferences
94- - Enable MCP server connections in Settings → Developer → MCP Servers
95-
96- ## MCP servers and extensions
97-
98- Model Context Protocol (MCP) servers extend Claude's capabilities by providing access to external
99- tools, APIs, and data sources. They enable Claude to interact with your development environment,
100- databases, and other services while maintaining security boundaries.
101-
102- ### Bitwarden AI plugin marketplace
103-
104- Bitwarden maintains a curated [ marketplace of AI plugins] ( https://github.com/bitwarden/ai-plugins )
105- specifically designed for our development workflows. This marketplace was created to provide
106- quality-controlled, security-reviewed plugins that follow Bitwarden's coding standards and security
107- requirements. All marketplace plugins are maintained by the Bitwarden team and include comprehensive
108- documentation, testing, and security validation.
109-
110- To use the marketplace with Claude Code:
111-
112- ``` bash
113- /plugin marketplace add bitwarden/ai-plugins
114- ```
115-
116- ### Understanding MCP servers
117-
118- MCP servers are separate processes that communicate with Claude through a standardized protocol.
119- They can:
120-
121- - Access local file systems and databases
122- - Execute commands and scripts
123- - Integrate with third-party APIs
124- - Provide specialized reasoning capabilities
125-
126- We recommend at least two be installed by everyone:
127-
128- ### Installing Sequential Thinking MCP server
129-
130- The Sequential Thinking server enhances Claude's problem-solving capabilities by providing
131- structured, step-by-step reasoning for complex tasks.
132-
133- #### Claude Code
134-
135- ``` bash
136- claude mcp add --scope user sequential-thinking -- npx -y @modelcontextprotocol/server-sequential-thinking
137- ```
138-
139- #### Claude Desktop
140-
141- Edit your ` ~/.claude.json ` , go to the ` mcpServers ` section and add:
142-
143- ``` json
144- "sequential-thinking" : {
145- "type" : " stdio" ,
146- "command" : " npx" ,
147- "args" : [
148- " -y" ,
149- " @modelcontextprotocol/server-sequential-thinking"
150- ]
151- }
152- ```
153-
154- Restart Claude Desktop to activate the server.
155-
156- ### Installing Memory MCP server
157-
158- The Memory server provides Claude with persistent memory capabilities, allowing it to remember
159- context across sessions and maintain a knowledge graph of your projects.
160-
161- #### Claude Code
162-
163- ``` bash
164- claude mcp add --scope user memory -- npx -y @modelcontextprotocol/server-memory
165- ```
166-
167- #### Claude Desktop
168-
169- Edit your ` ~/.claude.json ` , go to the ` mcpServers ` section and add:
170-
171- ``` json
172- "memory" : {
173- "type" : " stdio" ,
174- "command" : " npx" ,
175- "args" : [
176- " -y" ,
177- " @modelcontextprotocol/server-memory"
178- ]
179- }
180- ```
181-
182- Restart Claude Desktop to activate the server.
183-
184- ### Verifying installations
185-
186- #### Claude Code
187-
188- ``` bash
189- claude mcp list
190- ```
191-
192- #### Claude Desktop
193-
194- 1 . Open Claude Desktop
195- 2 . Start a new conversation
196- 3 . Type: "Can you list your available MCP servers?"
197- 4 . Claude should respond with the configured servers
198-
199- ### Troubleshooting
200-
201- Common issues and solutions:
202-
203- ** Server not starting** :
204-
205- - Verify NPM packages are installed globally
206- - Check Node version (must be 18+)
207- - Review server logs in ` ~/.claude-code/logs/ ` or Claude Desktop's developer console
208-
209- ** Permission errors** :
210-
211- - Ensure data directories have proper permissions
212- - On macOS/Linux: ` chmod 755 ~/.claude-memory `
213-
214- ** Configuration not loading** :
215-
216- - Validate JSON syntax in configuration files
217- - Restart Claude Code or Claude Desktop after configuration changes
218-
219- ### Best practices
220-
221- ** Security considerations** :
222-
223- - Only install MCP servers from trusted sources
224- - Review server permissions and capabilities before installation:
225- - Examine the server's source code or documentation to understand what file system access it
226- requires
227- - Verify what external APIs or services the server connects to
228- - Check if the server executes system commands and understand which ones
229- - Confirm whether the server stores persistent data and where it's stored
230- - Review network permissions and ensure the server only communicates with expected endpoints
231- - Validate that the server follows principle of least privilege
232- - Use trusted LLM providers and models:
233- - Prefer established providers with strong security track records (e.g., Anthropic)
234- - Verify the provider's data handling policies and ensure they align with Bitwarden's security
235- requirements
236- - Confirm that your API keys and credentials are stored securely
237- - Understand whether your prompts and code are used for model training (opt out if possible)
238- - Use enterprise or business tier services when available for enhanced security guarantees
239- - Core model usage guidelines:
240- - Use the latest stable model versions to benefit from security improvements and bug fixes
241- - Avoid deprecated or experimental models in production workflows
242- - Be aware of model capabilities and limitations - not all models are suitable for code generation
243- - Consider model context windows and token limits when designing workflows
244- - Use model-specific features (like Claude's extended thinking) appropriately for complex tasks
245- - Monitor model output for hallucinations or incorrect information, especially in
246- security-critical code
247- - Regularly update servers to get security patches
248-
249- ** Performance optimization** :
250-
251- - Limit the number of active servers to those you actively use
252- - Monitor resource usage, especially for memory-intensive servers
253- - Configure appropriate timeouts for long-running operations
254-
255- ** Data management** :
256-
257- - Regularly backup memory server data directories
258- - Clear old session data periodically to maintain performance
259- - Use project-specific memory contexts when appropriate
260-
261- ** Integration with development workflow** :
262-
263- - Configure project-specific MCP servers in repository ` .claude/ ` directories
264- - Document custom MCP server requirements in project README files
265- - Share MCP configurations with team members for consistency
56+ :::
0 commit comments