frontend_design_loop_eval
Evaluates frontend design changes by applying patches, running tests/linting, capturing screenshots, and using vision AI to score visual quality and detect issues.
Instructions
MCP tool wrapper for _frontend_design_loop_eval_impl.
Returns:
JSON summary (TextContent)
Optional vision instructions (TextContent) when vision_provider=client
Optional screenshots as ImageContent (base64) so Claude can use built-in vision
Tool call template
Replace each placeholder with a value matching the input schema. Only required parameters are included; add optional parameters as needed.
{
"id": 1,
"jsonrpc": "2.0",
"method": "tools/call",
"params": {
"arguments": {
"repo_path": "<repo_path>",
"patches": "<patches>"
},
"name": "frontend_design_loop_eval"
}
}Input Schema
| Name | Required | Description | Default |
|---|---|---|---|
| goal | No | ||
| patches | Yes | ||
| repo_path | Yes | ||
| viewports | No | ||
| preview_url | No | ||
| vision_mode | No | auto | |
| lint_command | No | ||
| test_command | No | ||
| vision_model | No | gemini-2.0-flash | |
| keep_worktree | No | ||
| include_images | No | ||
| gate_timeout_ms | No | ||
| preview_command | No | ||
| vision_provider | No | client | |
| worktree_reuse_dirs | No | ||
| unsafe_shell_commands | No | ||
| preview_wait_timeout_s | No | ||
| vision_score_threshold | No | ||
| unsafe_external_preview | No | ||
| include_vision_instructions | No | ||
| vision_broken_min_confidence | No |
Output Schema
| Name | Required | Description | Default |
|---|---|---|---|
| result | Yes |