refactor(core): upgrade ai-sdk v4 → v6 + Anthropic/openai-compatible reasoning round-trip + max-thinking budget audit

Migrates the LLM stack from [email protected] + @ai-sdk/[email protected] +
@ai-sdk/[email protected] to [email protected] + @ai-sdk/[email protected]
+ @ai-sdk/[email protected]. Full design in plan-v6-upgrade.md;
two rounds of Gemini code review captured in report.md.

Motivation: the recurring 'reasoning-signature without reasoning' error
on Claude Opus 4.7 was a v4 SDK artefact — @ai-sdk/[email protected] emitted
Anthropic signature_delta as a separate stream chunk that orphaned when
the model produced a signed-but-empty thinking block, and our chunk
store had no signature field so the round-trip back to Anthropic was
rejected on the next turn. In v6, signatures arrive inside
providerMetadata on the reasoning-end event, and the orphan-signature
class of bug is gone at the SDK level.

Core changes:
  • ThinkingChunk gains optional metadata?: Record<string, unknown>
    (the v6 providerMetadata blob). A non-undefined metadata 'seals'
    the chunk: subsequent reasoning-delta opens a new chunk rather
    than extending the sealed one.
  • AgentEvent gains { type: 'reasoning-end'; metadata? } (replaces
    the v4 reasoning-signature variant).
  • toModelMessages (replaces toCoreMessages):
      - returns ModelMessage[] (was CoreMessage[])
      - thinking → { type: 'reasoning', text, providerOptions: metadata }
      - tool-batch entries → { type: 'tool-call', input } (was 'args')
      - tool results → { output: { type: 'text', value } } ToolResultOutput
  • Claude OAuth uses createAnthropic({ authToken }) natively — no more
    custom-fetch x-api-key → Bearer swap.
  • rewriteBodyForOpus47 deleted — Opus 4.7 adaptive thinking is native
    via providerOptions.anthropic.thinking = { type: 'adaptive' }.
  • V1 middleware → V3 (specificationVersion: 'v3').
  • v4-era normalizeMessages openai-compatible middleware deleted; the
    v6 openai-compatible provider extracts reasoning_content natively
    from { type: 'reasoning' } content parts.
  • applyAnthropicStructuralNormalisations (mirrors opencode
    provider/transform.ts:53-148): drops empty text/reasoning parts,
    scrubs non-[a-zA-Z0-9_-] toolCallIds, splits [tool-call, non-tool]
    assistant turns (Anthropic rejects tool_use followed by text).
  • applyOpenAICompatibleReasoningNormalisation (mirrors opencode
    transform.ts:217-249): lifts reasoning text into
    providerOptions.openaiCompatible.reasoning_content (always, even
    empty). Solves DeepSeek 'The reasoning_content in the thinking
    mode must be passed back' — the v6 SDK skips emitting
    reasoning_content when text is empty (dist/index.mjs:245), but
    DeepSeek requires the field present once thinking was used.
  • Tools: tool({ inputSchema: jsonSchema(zodToJsonSchema(...)) })
    (was parameters: ZodSchema). AI SDK tools have no execute
    callback — the agent runs tools manually for permission prompts
    and shell-output streaming. New dep: zod-to-json-schema@^3.25.2.
  • fullStream event loop rewritten for v6 event shape: text-delta
    (text not textDelta), reasoning-start/delta/end, tool-input-*,
    tool-call (input not args), tool-result, tool-error (new), abort
    (new), start-step/finish-step, finish.

Max-thinking audit (matches opencode transform.ts:642-671 budgets):
  • Claude enabled-thinking max budget 16000 → 31999 (Anthropic ceiling)
  • Claude enabled-thinking high budget 10000 → 16000
  • maxOutputTokens 'budget + 8000' → fixed 32000 (matches opencode's
    OUTPUT_TOKEN_MAX; model self-allocates thinking vs response within)
  • Opus 4.7 adaptive thinking gains display: 'summarized' and sibling
    effort field (without these, thinking content is hidden by Anthropic
    and the model barely thinks).

Frontend mirrors:
  • types.ts — ThinkingChunk.metadata?, AgentEvent reasoning-end
  • tabs.svelte.ts — routes reasoning-end through applyChunkEvent
  • ChatMessage.svelte — hides empty thinking chunks; hides the entire
    assistant bubble when no chunk has renderable content

Gemini-review-driven fixes:
  • tool-error and abort stream events now surface as error chunks
    (were silently ignored)
  • toolCallId scrubbing pass (opencode transform.ts:96-122 parity)
  • Empty-reasoning-cull explicit test coverage for both Anthropic
    structural normalisation and DeepSeek path

Test counts (223 tests across 3 packages, all green):
  • tests/chunks/append.test.ts: 44 (was 38) — reasoning-end sealing,
    orphan walk-back, multi-block interleaving
  • tests/agent/agent.test.ts: 24 (was 5) — exhaustive v6 event
    mappings, structural normalisations, signature/reasoning_content
    round-trip, tool-error/abort branches, DeepSeek scenario, empty
    reasoning edge case
  • tests/llm/provider.test.ts: 9 (was 22) — dropped 13 obsolete v4
    middleware tests; new minimal tests confirm no middleware wrapping
    on default openai-compat path and that createAnthropic gets
    authToken vs apiKey correctly for OAuth vs api-key flows
  • tests/tools/registry.test.ts: 10 (was 4) — v6 tool() contract
    (inputSchema, no execute, JSON Schema for nested zod)
  • packages/api/tests/agent-manager.test.ts: 12 (was 7) — mock Agent
    emits v6 reasoning events; reasoning-end broadcast + ordering
  • packages/frontend/tests/chat-store.test.ts: 35 (was 32) —
    reasoning-end flow through Svelte $state store

typecheck clean (tsc --noEmit on core + api, svelte-check on frontend),
biome clean across 124 files.

1 files changed, 101 insertions, 8 deletions

diff --git a/packages/core/tests/tools/registry.test.ts b/packages/core/tests/tools/registry.test.ts
index b6f1fca..cad75d2 100644
--- a/packages/core/tests/tools/registry.test.ts
+++ b/packages/core/tests/tools/registry.test.ts
@@ -17,6 +17,21 @@ const anotherTool: ToolDefinition = {
 	execute: async (_args) => "another result",
 };
 
+/** A non-trivial tool that exercises nested objects, required fields, and enums. */
+const complexTool: ToolDefinition = {
+	name: "complex_tool",
+	description: "A tool with nested parameters",
+	parameters: z.object({
+		command: z.string().describe("Shell command to run"),
+		options: z.object({
+			timeout: z.number().optional().describe("Timeout in milliseconds"),
+			shell: z.enum(["bash", "sh", "zsh"]).describe("Shell to use"),
+		}),
+		flags: z.array(z.string()).optional().describe("Additional flags"),
+	}),
+	execute: async (_args) => "complex result",
+};
+
 describe("createToolRegistry", () => {
 	it("returns all tools via getTools()", () => {
 		const registry = createToolRegistry([mockTool, anotherTool]);
@@ -38,13 +53,91 @@ describe("createToolRegistry", () => {
 		expect(registry.getTool("nonexistent")).toBeUndefined();
 	});
 
-	it("getAISDKTools returns correct format", () => {
-		const registry = createToolRegistry([mockTool, anotherTool]);
-		const aiTools = registry.getAISDKTools();
-		expect(aiTools).toHaveProperty("mock_tool");
-		expect(aiTools).toHaveProperty("another_tool");
-		// Each should have description and parameters (AI SDK tool format)
-		expect(aiTools.mock_tool).toHaveProperty("description");
-		expect(aiTools.mock_tool).toHaveProperty("parameters");
+	describe("getAISDKTools", () => {
+		it("returns correct keys for all tools", () => {
+			const registry = createToolRegistry([mockTool, anotherTool]);
+			const aiTools = registry.getAISDKTools();
+			expect(aiTools).toHaveProperty("mock_tool");
+			expect(aiTools).toHaveProperty("another_tool");
+		});
+
+		it("AI SDK tools have description from ToolDefinition", () => {
+			const registry = createToolRegistry([mockTool]);
+			const aiTools = registry.getAISDKTools();
+			expect(aiTools.mock_tool.description).toBe("A mock tool for testing");
+		});
+
+		it("AI SDK tools surface schema via inputSchema, not parameters", () => {
+			const registry = createToolRegistry([mockTool]);
+			const aiTools = registry.getAISDKTools();
+			// v6 uses inputSchema; v4 used parameters — this verifies the migration
+			expect(aiTools.mock_tool).toHaveProperty("inputSchema");
+			expect(aiTools.mock_tool).not.toHaveProperty("parameters");
+		});
+
+		it("AI SDK tools have no execute callback so the SDK does not auto-run", () => {
+			const registry = createToolRegistry([mockTool, anotherTool, complexTool]);
+			const aiTools = registry.getAISDKTools();
+			for (const [name, sdkTool] of Object.entries(aiTools)) {
+				expect(
+					(sdkTool as Record<string, unknown>).execute,
+					`Tool "${name}" should not have an execute callback`,
+				).toBeUndefined();
+			}
+		});
+
+		it("inputSchema produces valid JSONSchema7 for a simple tool", () => {
+			const registry = createToolRegistry([mockTool]);
+			const aiTools = registry.getAISDKTools();
+			const schema = aiTools.mock_tool.inputSchema;
+			// jsonSchema() wraps the raw JSONSchema7; it should expose the schema
+			// as a `jsonSchema` property on the Schema object
+			expect(schema).toBeDefined();
+			// The wrapped schema object should carry the JSON Schema definition
+			const schemaObj = schema as { jsonSchema: Record<string, unknown> };
+			expect(schemaObj.jsonSchema).toBeDefined();
+			expect(schemaObj.jsonSchema.type).toBe("object");
+			const props = schemaObj.jsonSchema.properties as Record<string, unknown>;
+			expect(props).toHaveProperty("input");
+		});
+
+		it("inputSchema produces correct JSONSchema7 for a non-trivial nested tool", () => {
+			const registry = createToolRegistry([complexTool]);
+			const aiTools = registry.getAISDKTools();
+			const schema = aiTools.complex_tool.inputSchema;
+			expect(schema).toBeDefined();
+			const schemaObj = schema as { jsonSchema: Record<string, unknown> };
+			expect(schemaObj.jsonSchema.type).toBe("object");
+
+			const props = schemaObj.jsonSchema.properties as Record<string, Record<string, unknown>>;
+
+			// Top-level required field "command"
+			expect(props).toHaveProperty("command");
+			expect(props.command.type).toBe("string");
+
+			// Nested object "options"
+			expect(props).toHaveProperty("options");
+			expect(props.options.type).toBe("object");
+			const optProps = props.options.properties as Record<string, Record<string, unknown>>;
+			expect(optProps).toHaveProperty("shell");
+			expect(optProps.shell.enum).toEqual(["bash", "sh", "zsh"]);
+
+			// Optional array "flags" present as a property
+			expect(props).toHaveProperty("flags");
+			expect(props.flags.type).toBe("array");
+
+			// Required fields should include "command" and "options"
+			const required = schemaObj.jsonSchema.required as string[];
+			expect(required).toContain("command");
+			expect(required).toContain("options");
+		});
+
+		it("getTool still returns the original ToolDefinition with execute", () => {
+			const registry = createToolRegistry([mockTool]);
+			const def = registry.getTool("mock_tool");
+			expect(def).toBeDefined();
+			expect(typeof def?.execute).toBe("function");
+			expect(def?.name).toBe("mock_tool");
+		});
 	});
 });