You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Download the `.vsix` extension from the [releases page](https://github.com/docker/labs-ai-tools-vscode/releases).
10
+
[Configure Docker Desktop](../docs/claude-desktop) to use the `mcp/run` Docker container.
11
11
12
-
### Install extension from `.vsix` file
12
+
### Restart Claude Desktop
13
13
14
-
```sh
15
-
code --install-extension <path-to-vsix-file>
16
-
```
14
+
Restarting desktop should be a one-time activity. However, Claude
15
+
Desktop does not support the `tools/list_changed` notification so we
16
+
currently have to restart desktop more less continuously. :)
17
17
18
-
or use the VSCode command palette `Extensions: Install from VSIX...`
18
+
### Try a prompt
19
19
20
-
### Open a prompt
21
-
With the extension installed, open a prompt in VSCode. Examples can be found in the [examples](https://github.com/docker/labs-ai-tools-for-devs/tree/main/prompts/examples) directory.
20
+
Write a prompt in Claude that will run one of the tools in a registered defintion.
21
+
For example:
22
22
23
-
### Configure OpenAI API Key, or use a different model.
24
-
Default model is `gpt-4` provided by OpenAI. Use an the VSCode command palette `Docker AI: Set secret key` to set your API key.
25
-
26
-
**Changing the model:**
27
-
Use the following keys in the prompt front-matter to change the model:
28
-
29
-
Anthropic:
30
-
```yml
31
-
---
32
-
url: https://api.anthropic.com
33
-
model: claude-3-5-sonnet-20240620
34
-
---
35
-
```
36
-
37
-
38
-
Ollama:
39
-
```yml
40
-
---
41
-
model: llama3.2
42
-
url: https://docker.host.internal:11434/v1
43
-
---
44
-
```
45
-
46
-
### Run the prompt
47
-
Use the VSCode command palette `Docker AI: Run this prompt` to run the prompt.
23
+
> Use hello world to send a greeting and then respond to what comes back.
Download the `.vsix` extension from the [releases page](https://github.com/docker/labs-ai-tools-vscode/releases).
11
+
12
+
### Install extension from `.vsix` file
13
+
14
+
```sh
15
+
code --install-extension <path-to-vsix-file>
16
+
```
17
+
18
+
or use the VSCode command palette `Extensions: Install from VSIX...`
19
+
20
+
### Open a prompt
21
+
With the extension installed, open a prompt in VSCode. Examples can be found in the [examples](https://github.com/docker/labs-ai-tools-for-devs/tree/main/prompts/examples) directory.
22
+
23
+
### Configure OpenAI API Key, or use a different model.
24
+
Default model is `gpt-4` provided by OpenAI. Use an the VSCode command palette `Docker AI: Set secret key` to set your API key.
25
+
26
+
**Changing the model:**
27
+
Use the following keys in the prompt front-matter to change the model:
28
+
29
+
Anthropic:
30
+
```yml
31
+
---
32
+
model: claude-3-5-sonnet-20240620
33
+
---
34
+
```
35
+
36
+
37
+
Ollama:
38
+
```yml
39
+
---
40
+
model: llama3.2
41
+
url: https://docker.host.internal:11434/v1
42
+
---
43
+
```
44
+
45
+
### Run the prompt
46
+
Use the VSCode command palette `Docker AI: Run this prompt` to run the prompt.
0 commit comments