plan_feature
Generate a structured implementation plan for a UI feature from natural language intent.
Reference
Parameters
- Name
intent- Type
- string · required
- Description
Natural language description of the feature to build (min 1)
- Name
context- Type
- object · optional
- Description
Contextual information for better planning
- Name
options- Type
- object · optional
- Description
Planning options
Description
Generate a structured implementation plan for a UI feature from natural language intent.
Returns an ordered plan with presets, components, validation checkpoints, and constraints. Composes data from recipes, intent routes, context rules, and presets — no LLM calls.
USE THIS when building multi-component features (e.g., "settings page with profile editing"). Use get_preferred_component for single components, get_workflow for tool orchestration.
Parameters:
- intent (required): Natural language feature description (e.g., "login form with email and password")
- context (optional): { filePath, location, existingComponents[] }
- options (optional): { maxSteps (default 15), includeCode (default true), includeFileStructure (default true), layerId }
Returns ordered steps with phase, component, preset, code, imports, props, and dependencies. Also includes validation checkpoints, constraints, anti-patterns, and confidence score (0.9 = recipe match, 0.7 = DB route match, 0.3 = core fallback).
Source
Defined in packages/mcp/src/tools.ts.
The request handler lives alongside its registration in packages/mcp/src/tools/.