refactor(model-core): remove src back-imports via core utilities and adapter

- move fuzzyMatchModel and transformModelForProvider into model-core

- replace connected-providers re-export with adapter contract

- route shared wrappers through model-core exports

Ultraworked with [Sisyphus](https://github.com/code-yeongyu/oh-my-openagent)

Co-authored-by: Sisyphus <clio-agent@sisyphuslabs.ai>

src/shared/model-availability.ts

@@ -5,6 +5,12 @@ import { getOpenCodeCacheDir } from "./data-path"
 import * as connectedProvidersCache from "./connected-providers-cache"
 import { normalizeSDKResponse } from "./normalize-sdk-response"
 
+function normalizeModelName(name: string): string {
+	return name
+		.toLowerCase()
+		.replace(/claude-(opus|sonnet|haiku)-(\d+)[.-](\d+)/g, "claude-$1-$2.$3")
+}
+
 /**
  * Fuzzy match a target model name against available models
  * 
@@ -25,12 +31,6 @@ import { normalizeSDKResponse } from "./normalize-sdk-response"
  * fuzzyMatchModel("gpt-5.4", available) // → "openai/gpt-5.4"
  * fuzzyMatchModel("claude", available, ["openai"]) // → null (provider filter excludes anthropic)
  */
-function normalizeModelName(name: string): string {
-	return name
-		.toLowerCase()
-		.replace(/claude-(opus|sonnet|haiku)-(\d+)[.-](\d+)/g, "claude-$1-$2.$3")
-}
-
 export function fuzzyMatchModel(
 	target: string,
 	available: Set<string>,
@@ -45,7 +45,6 @@ export function fuzzyMatchModel(
 
 	const targetNormalized = normalizeModelName(target)
 
-	// Filter by providers if specified
 	let candidates = Array.from(available)
 	if (providers && providers.length > 0) {
 		const providerSet = new Set(providers)
@@ -61,7 +60,6 @@ export function fuzzyMatchModel(
 		return null
 	}
 
-	// Find all matches (case-insensitive substring match with normalization)
 	const matches = candidates.filter((model) =>
 		normalizeModelName(model).includes(targetNormalized),
 	)
@@ -73,16 +71,12 @@ export function fuzzyMatchModel(
 		return null
 	}
 
-	// Priority 1: Exact match (normalized full model string)
 	const exactMatch = matches.find((model) => normalizeModelName(model) === targetNormalized)
 	if (exactMatch) {
 		log("[fuzzyMatchModel] exact match found", { exactMatch })
 		return exactMatch
 	}
 
-	// Priority 2: Exact model ID match (part after provider/)
-	// This ensures "big-pickle" matches "zai-coding-plan/big-pickle" over "zai-coding-plan/glm-5"
-	// Use filter + shortest to handle multi-provider cases (e.g., openai/gpt-5.4 + opencode/gpt-5.4)
 	const exactModelIdMatches = matches.filter((model) => {
 		const modelId = model.split("/").slice(1).join("/")
 		return normalizeModelName(modelId) === targetNormalized
@@ -95,7 +89,6 @@ export function fuzzyMatchModel(
 		return result
 	}
 
-	// Priority 3: Shorter model name (more specific, fallback for partial matches)
 	const result = matches.reduce((shortest, current) =>
 		current.length < shortest.length ? current : shortest,
 	)
@@ -103,13 +96,6 @@ export function fuzzyMatchModel(
 	return result
 }
 
-/**
- * Check if a target model is available (fuzzy match by model name, no provider filtering)
- * 
- * @param targetModel - Model name to check (e.g., "gpt-5.3-codex")
- * @param availableModels - Set of available models in "provider/model" format
- * @returns true if model is available, false otherwise
- */
 export function isModelAvailable(
 	targetModel: string,
 	availableModels: Set<string>,

src/shared/model-resolver.ts

@@ -1,12 +1,28 @@
+import {
+	resolveModel,
+	resolveModelWithFallback as resolveModelWithFallbackFromCore,
+	normalizeFallbackModels,
+	flattenToFallbackModelStrings,
+} from "@oh-my-opencode/model-core"
+import type {
+	ModelResolutionInput,
+	ExtendedModelResolutionInput,
+} from "@oh-my-opencode/model-core"
+import * as connectedProvidersCache from "./connected-providers-cache"
+
+export { resolveModel, normalizeFallbackModels, flattenToFallbackModelStrings }
+
+type CoreModelResolutionResult = ReturnType<typeof resolveModelWithFallbackFromCore>
+export type ModelResolutionResult = Exclude<CoreModelResolutionResult, undefined>
+export type ModelSource = ModelResolutionResult["source"]
+
+export function resolveModelWithFallback(
+	input: ExtendedModelResolutionInput,
+): CoreModelResolutionResult {
+	return resolveModelWithFallbackFromCore(input, connectedProvidersCache)
+}
+
 export type {
-  ModelResolutionInput,
-  ModelSource,
-  ModelResolutionResult,
-  ExtendedModelResolutionInput,
-} from "@oh-my-opencode/model-core"
-export {
-  resolveModel,
-  resolveModelWithFallback,
-  normalizeFallbackModels,
-  flattenToFallbackModelStrings,
-} from "@oh-my-opencode/model-core"
+	ModelResolutionInput,
+	ExtendedModelResolutionInput,
+}

src/shared/provider-model-id-transform.ts

@@ -23,18 +23,13 @@ function applyGatewayTransforms(model: string): string {
 }
 
 export function transformModelForProvider(provider: string, model: string): string {
-	// Vercel AI Gateway expects <sub-provider>/<model> (e.g. anthropic/claude-opus-4.7).
-	// Canonical names in model-requirements.ts may be bare (claude-opus-4-7) or
-	// already prefixed (anthropic/claude-opus-4-7). Both need gateway-specific transforms.
 	if (provider === "vercel") {
-		// Already prefixed — transform only the model part
 		const slashIndex = model.indexOf("/")
 		if (slashIndex !== -1) {
 			const subProvider = model.substring(0, slashIndex)
 			const subModel = model.substring(slashIndex + 1)
 			return `${subProvider}/${applyGatewayTransforms(subModel)}`
 		}
-		// Bare name — infer sub-provider from model prefix (claude- → anthropic, etc.)
 		const subProvider = inferSubProvider(model)
 		if (subProvider) {
 			return `${subProvider}/${applyGatewayTransforms(model)}`