@vizejs/musea-mcp-server 0.61.0 → 0.63.0
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/README.md +35 -4
- package/dist/cli.mjs +1 -1
- package/dist/index.mjs +1 -1
- package/dist/src-Cp5GZpzj.mjs +1840 -0
- package/package.json +2 -2
- package/dist/src-DqveCZc0.mjs +0 -780
package/README.md
CHANGED
|
@@ -42,10 +42,41 @@ vp dlx @vizejs/musea-mcp-server --project ./my-vue-app
|
|
|
42
42
|
|
|
43
43
|
## MCP Tools
|
|
44
44
|
|
|
45
|
-
- `
|
|
46
|
-
- `
|
|
47
|
-
- `
|
|
48
|
-
- `
|
|
45
|
+
- `analyze_component` - Extract props and emits from a Vue SFC or linked art component
|
|
46
|
+
- `get_palette` - Derive an interactive props palette from an art file
|
|
47
|
+
- `list_components` - List registered components with status, variants, and resource URIs
|
|
48
|
+
- `get_component` - Get full component details, analysis, palette, and docs
|
|
49
|
+
- `get_variant` - Get one specific variant
|
|
50
|
+
- `search_components` - Search components by title, tags, category, component name, and variants
|
|
51
|
+
- `recommend_components` - Rank components by user intent or UI task
|
|
52
|
+
- `generate_variants` - Generate an `.art.vue` draft from a component
|
|
53
|
+
- `generate_csf` - Convert an art file to Storybook CSF
|
|
54
|
+
- `generate_docs` - Generate Markdown docs for a component
|
|
55
|
+
- `generate_catalog` - Generate a catalog for the whole design system
|
|
56
|
+
- `get_tokens` - Read design tokens
|
|
57
|
+
- `search_tokens` - Search tokens without loading the full token tree
|
|
58
|
+
|
|
59
|
+
## Reproducible Prompting
|
|
60
|
+
|
|
61
|
+
Musea helps AI assistants stay grounded in real component data, but reusable prompts still need to
|
|
62
|
+
be tuned for reproducibility.
|
|
63
|
+
|
|
64
|
+
A practical loop is:
|
|
65
|
+
|
|
66
|
+
1. Fix a few evaluation scenarios first.
|
|
67
|
+
2. Run the prompt in fresh assistant sessions.
|
|
68
|
+
3. Collect both self-reported ambiguity and tool/checklist traces.
|
|
69
|
+
4. Fix one ambiguity per iteration.
|
|
70
|
+
|
|
71
|
+
For Musea-specific workflows, check whether the assistant:
|
|
72
|
+
|
|
73
|
+
- used real components from the registry
|
|
74
|
+
- matched prop and variant names to MCP output
|
|
75
|
+
- read tokens before suggesting visual values
|
|
76
|
+
- reported missing metadata instead of guessing
|
|
77
|
+
|
|
78
|
+
The full MCP guide includes a longer reproducibility workflow and prompt template:
|
|
79
|
+
[docs/content/integrations/mcp.md](https://github.com/ubugeeei/vize/blob/main/docs/content/integrations/mcp.md)
|
|
49
80
|
|
|
50
81
|
## License
|
|
51
82
|
|
package/dist/cli.mjs
CHANGED
package/dist/index.mjs
CHANGED
|
@@ -1,2 +1,2 @@
|
|
|
1
|
-
import { n as startServer, t as createMuseaServer } from "./src-
|
|
1
|
+
import { n as startServer, t as createMuseaServer } from "./src-Cp5GZpzj.mjs";
|
|
2
2
|
export { createMuseaServer, createMuseaServer as default, startServer };
|