Vendor OpenClaw source as Adolf fork baseline

Adolf is a fork/vendored clone of github.com/openclaw/openclaw (v2026.6.11),
free to diverge. Tree copied sans upstream .git; upstream remote added for
future syncs. Node pinned to 24 (.nvmrc); engines already require >=22.19.
Preserves docs/ARCHITECTURE.md.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01LeqyaxJF2nbRXJtae2kNB2

extensions/amazon-bedrock-mantle/discovery.test.ts

@@ -0,0 +1,693 @@
+// Amazon Bedrock Mantle tests cover discovery plugin behavior.
+import { afterEach, beforeEach, describe, expect, it, vi } from "vitest";
+
+const {
+  discoverMantleModels,
+  generateBearerTokenFromIam,
+  getCachedIamToken,
+  MANTLE_IAM_TOKEN_MARKER,
+  mergeImplicitMantleProvider,
+  resetIamTokenCacheForTest,
+  resetMantleDiscoveryCacheForTest,
+  resolveImplicitMantleProvider,
+  resolveMantleBearerToken,
+  resolveMantleRuntimeBearerToken,
+} = await import("./api.js");
+
+function createTokenProviderFactory(tokenProvider: () => Promise<string>) {
+  return vi.fn(() => tokenProvider);
+}
+
+type MockWithCalls = {
+  mock: { calls: unknown[][] };
+};
+
+function argAt(mock: MockWithCalls, callIndex: number, argIndex: number): unknown {
+  const call = mock.mock.calls[callIndex];
+  if (!call) {
+    throw new Error(`expected call ${callIndex}`);
+  }
+  if (!(argIndex in call)) {
+    throw new Error(`expected call ${callIndex} argument ${argIndex}`);
+  }
+  return call[argIndex];
+}
+
+function objectArgAt(
+  mock: MockWithCalls,
+  callIndex: number,
+  argIndex: number,
+): Record<string, unknown> {
+  const value = argAt(mock, callIndex, argIndex);
+  if (value === undefined || value === null || typeof value !== "object" || Array.isArray(value)) {
+    throw new Error(`expected call ${callIndex} argument ${argIndex} to be an object`);
+  }
+  return value as Record<string, unknown>;
+}
+
+function stringArgAt(mock: MockWithCalls, callIndex: number, argIndex: number): string {
+  const value = argAt(mock, callIndex, argIndex);
+  if (typeof value !== "string") {
+    throw new Error(`expected call ${callIndex} argument ${argIndex} to be a string`);
+  }
+  return value;
+}
+
+function recordField(value: unknown, field: string): Record<string, unknown> {
+  if (value === undefined || value === null || typeof value !== "object" || Array.isArray(value)) {
+    throw new Error(`expected ${field} to be an object`);
+  }
+  return value as Record<string, unknown>;
+}
+
+describe("bedrock mantle discovery", () => {
+  const originalEnv = process.env;
+
+  beforeEach(() => {
+    process.env = { ...originalEnv };
+    vi.restoreAllMocks();
+    resetMantleDiscoveryCacheForTest();
+    resetIamTokenCacheForTest();
+  });
+
+  afterEach(() => {
+    vi.restoreAllMocks();
+    resetMantleDiscoveryCacheForTest();
+    resetIamTokenCacheForTest();
+    process.env = originalEnv;
+  });
+
+  // ---------------------------------------------------------------------------
+  // Bearer token resolution
+  // ---------------------------------------------------------------------------
+
+  it("resolves bearer token from AWS_BEARER_TOKEN_BEDROCK", () => {
+    expect(
+      resolveMantleBearerToken({
+        AWS_BEARER_TOKEN_BEDROCK: "bedrock-api-key-abc123", // pragma: allowlist secret
+      } as NodeJS.ProcessEnv),
+    ).toBe("bedrock-api-key-abc123");
+  });
+
+  it("returns undefined when no bearer token env var is set", () => {
+    expect(resolveMantleBearerToken({} as NodeJS.ProcessEnv)).toBeUndefined();
+  });
+
+  it("trims whitespace from bearer token", () => {
+    expect(
+      resolveMantleBearerToken({
+        AWS_BEARER_TOKEN_BEDROCK: "  my-token  ", // pragma: allowlist secret
+      } as NodeJS.ProcessEnv),
+    ).toBe("my-token");
+  });
+
+  // ---------------------------------------------------------------------------
+  // IAM token generation
+  // ---------------------------------------------------------------------------
+
+  it("generates token from IAM credentials when token generation succeeds", async () => {
+    const tokenProvider = vi.fn(async () => "bedrock-api-key-generated"); // pragma: allowlist secret
+    const tokenProviderFactory = createTokenProviderFactory(tokenProvider);
+
+    const token = await generateBearerTokenFromIam({
+      region: "us-east-1",
+      tokenProviderFactory,
+    });
+
+    expect(token).toBe("bedrock-api-key-generated");
+    expect(tokenProviderFactory).toHaveBeenCalledWith({
+      region: "us-east-1",
+      expiresInSeconds: 7200,
+    });
+    expect(tokenProvider).toHaveBeenCalledTimes(1);
+  });
+
+  it("caches generated IAM tokens within TTL", async () => {
+    const tokenProvider = vi.fn(async () => "bedrock-api-key-cached"); // pragma: allowlist secret
+    const tokenProviderFactory = createTokenProviderFactory(tokenProvider);
+    let now = 1000;
+
+    const t1 = await generateBearerTokenFromIam({
+      region: "us-east-1",
+      now: () => now,
+      tokenProviderFactory,
+    });
+    now += 1800_000; // 30 min — within 2hr cache TTL
+    const t2 = await generateBearerTokenFromIam({
+      region: "us-east-1",
+      now: () => now,
+      tokenProviderFactory,
+    });
+
+    expect(t1).toEqual(t2);
+    expect(tokenProvider).toHaveBeenCalledTimes(1);
+  });
+
+  it("does not reuse an IAM token across regions", async () => {
+    const tokenProvider = vi
+      .fn<() => Promise<string>>()
+      .mockResolvedValueOnce("bedrock-api-key-east") // pragma: allowlist secret
+      .mockResolvedValueOnce("bedrock-api-key-west"); // pragma: allowlist secret
+    const tokenProviderFactory = createTokenProviderFactory(tokenProvider);
+
+    const east = await generateBearerTokenFromIam({
+      region: "us-east-1",
+      now: () => 1000,
+      tokenProviderFactory,
+    });
+    const west = await generateBearerTokenFromIam({
+      region: "us-west-2",
+      now: () => 2000,
+      tokenProviderFactory,
+    });
+
+    expect(east).toBe("bedrock-api-key-east");
+    expect(west).toBe("bedrock-api-key-west");
+    expect(tokenProviderFactory).toHaveBeenNthCalledWith(1, {
+      region: "us-east-1",
+      expiresInSeconds: 7200,
+    });
+    expect(tokenProviderFactory).toHaveBeenNthCalledWith(2, {
+      region: "us-west-2",
+      expiresInSeconds: 7200,
+    });
+    expect(tokenProvider).toHaveBeenCalledTimes(2);
+  });
+
+  it("returns undefined when IAM token generation fails", async () => {
+    const tokenProviderFactory = vi.fn(() => {
+      throw new Error("no credentials");
+    });
+
+    await expect(
+      generateBearerTokenFromIam({ region: "us-east-1", tokenProviderFactory }),
+    ).resolves.toBeUndefined();
+  });
+
+  it("skips IAM token generation when plugin discovery is disabled", async () => {
+    const tokenProviderFactory = vi.fn(() => {
+      throw new Error("disabled discovery should not generate a token");
+    });
+
+    await expect(
+      resolveImplicitMantleProvider({
+        env: { AWS_REGION: "us-east-1" } as NodeJS.ProcessEnv,
+        pluginConfig: { discovery: { enabled: false } },
+        tokenProviderFactory,
+      }),
+    ).resolves.toBeNull();
+
+    expect(tokenProviderFactory).not.toHaveBeenCalled();
+  });
+
+  it("getCachedIamToken returns cached token when valid", async () => {
+    const tokenProvider = vi.fn(async () => "bedrock-cached-token"); // pragma: allowlist secret
+    const tokenProviderFactory = createTokenProviderFactory(tokenProvider);
+
+    // Generate a token to populate the cache
+    await generateBearerTokenFromIam({ region: "us-east-1", tokenProviderFactory });
+
+    // Sync read should return the cached token
+    expect(getCachedIamToken("us-east-1")).toBe("bedrock-cached-token");
+  });
+
+  it("getCachedIamToken returns undefined when cache is empty", () => {
+    expect(getCachedIamToken("us-east-1")).toBeUndefined();
+  });
+
+  it("getCachedIamToken returns undefined when cache is expired", async () => {
+    const tokenProvider = vi.fn(async () => "bedrock-expired-token"); // pragma: allowlist secret
+    const tokenProviderFactory = createTokenProviderFactory(tokenProvider);
+
+    // Generate with a time far in the past so it's already expired
+    await generateBearerTokenFromIam({
+      region: "us-east-1",
+      now: () => 1000,
+      tokenProviderFactory,
+    });
+
+    // The cache entry exists but expiresAt is 1000 + 3600000 = 3601000
+    // Current Date.now() is way past that, so it should be expired
+    expect(getCachedIamToken("us-east-1")).toBeUndefined();
+  });
+
+  it("does not cache generated IAM tokens when ttl expiry overflows", async () => {
+    const tokenProvider = vi
+      .fn<() => Promise<string>>()
+      .mockResolvedValueOnce("bedrock-overflow-token-1") // pragma: allowlist secret
+      .mockResolvedValueOnce("bedrock-overflow-token-2"); // pragma: allowlist secret
+    const tokenProviderFactory = createTokenProviderFactory(tokenProvider);
+
+    await expect(
+      generateBearerTokenFromIam({
+        region: "us-east-1",
+        now: () => 8_640_000_000_000_000,
+        tokenProviderFactory,
+      }),
+    ).resolves.toBe("bedrock-overflow-token-1");
+    expect(getCachedIamToken("us-east-1")).toBeUndefined();
+
+    await expect(
+      generateBearerTokenFromIam({
+        region: "us-east-1",
+        now: () => 8_640_000_000_000_000,
+        tokenProviderFactory,
+      }),
+    ).resolves.toBe("bedrock-overflow-token-2");
+    expect(tokenProvider).toHaveBeenCalledTimes(2);
+  });
+
+  // ---------------------------------------------------------------------------
+  // Model discovery
+  // ---------------------------------------------------------------------------
+
+  it("discovers models from Mantle /v1/models endpoint sorted by id", async () => {
+    const mockFetch = vi.fn().mockResolvedValue({
+      ok: true,
+      json: async () => ({
+        data: [
+          { id: "openai.gpt-oss-120b", object: "model", owned_by: "openai" },
+          { id: "anthropic.claude-sonnet-4-6", object: "model", owned_by: "anthropic" },
+          { id: "mistral.devstral-2-123b", object: "model", owned_by: "mistral" },
+        ],
+      }),
+    });
+
+    const models = await discoverMantleModels({
+      region: "us-east-1",
+      bearerToken: "test-token",
+      fetchFn: mockFetch as unknown as typeof fetch,
+    });
+
+    expect(models).toHaveLength(3);
+    // Models should be sorted alphabetically by id
+    expect(models[0]?.id).toBe("anthropic.claude-sonnet-4-6");
+    expect(models[0]?.name).toBe("anthropic.claude-sonnet-4-6");
+    expect(models[0]?.reasoning).toBe(false);
+    expect(models[0]?.input).toEqual(["text"]);
+    expect(models[1]?.id).toBe("mistral.devstral-2-123b");
+    expect(models[1]?.reasoning).toBe(false);
+    expect(models[2]?.id).toBe("openai.gpt-oss-120b");
+    expect(models[2]?.reasoning).toBe(true); // GPT-OSS 120B supports reasoning
+
+    // Verify correct endpoint and auth header
+    expect(stringArgAt(mockFetch, 0, 0)).toBe("https://bedrock-mantle.us-east-1.api.aws/v1/models");
+    expect(recordField(objectArgAt(mockFetch, 0, 1).headers, "headers").Authorization).toBe(
+      "Bearer test-token",
+    );
+  });
+
+  it("infers reasoning support from model IDs", async () => {
+    const mockFetch = vi.fn().mockResolvedValue({
+      ok: true,
+      json: async () => ({
+        data: [
+          { id: "moonshotai.kimi-k2-thinking", object: "model" },
+          { id: "openai.gpt-oss-120b", object: "model" },
+          { id: "openai.gpt-oss-safeguard-120b", object: "model" },
+          { id: "deepseek.v3.2", object: "model" },
+          { id: "mistral.mistral-large-3-675b-instruct", object: "model" },
+        ],
+      }),
+    });
+
+    const models = await discoverMantleModels({
+      region: "us-east-1",
+      bearerToken: "test-token",
+      fetchFn: mockFetch as unknown as typeof fetch,
+    });
+
+    const byId = Object.fromEntries(models.map((m) => [m.id, m]));
+    expect(byId["moonshotai.kimi-k2-thinking"]?.reasoning).toBe(true);
+    expect(byId["openai.gpt-oss-120b"]?.reasoning).toBe(true);
+    expect(byId["openai.gpt-oss-safeguard-120b"]?.reasoning).toBe(true);
+    expect(byId["deepseek.v3.2"]?.reasoning).toBe(false);
+    expect(byId["mistral.mistral-large-3-675b-instruct"]?.reasoning).toBe(false);
+  });
+
+  it("returns empty array on permission error", async () => {
+    const mockFetch = vi.fn().mockResolvedValue({
+      ok: false,
+      status: 403,
+      statusText: "Forbidden",
+    });
+
+    const models = await discoverMantleModels({
+      region: "us-east-1",
+      bearerToken: "test-token",
+      fetchFn: mockFetch as unknown as typeof fetch,
+    });
+
+    expect(models).toStrictEqual([]);
+  });
+
+  it("returns empty array on network error", async () => {
+    const mockFetch = vi.fn().mockRejectedValue(new Error("ECONNREFUSED"));
+
+    const models = await discoverMantleModels({
+      region: "us-east-1",
+      bearerToken: "test-token",
+      fetchFn: mockFetch as unknown as typeof fetch,
+    });
+
+    expect(models).toStrictEqual([]);
+  });
+
+  it("filters out models with empty IDs", async () => {
+    const mockFetch = vi.fn().mockResolvedValue({
+      ok: true,
+      json: async () => ({
+        data: [
+          { id: "anthropic.claude-sonnet-4-6", object: "model" },
+          { id: "", object: "model" },
+          { id: "  ", object: "model" },
+        ],
+      }),
+    });
+
+    const models = await discoverMantleModels({
+      region: "us-east-1",
+      bearerToken: "test-token",
+      fetchFn: mockFetch as unknown as typeof fetch,
+    });
+
+    expect(models).toHaveLength(1);
+    expect(models[0]?.id).toBe("anthropic.claude-sonnet-4-6");
+  });
+
+  // ---------------------------------------------------------------------------
+  // Discovery caching
+  // ---------------------------------------------------------------------------
+
+  it("returns cached models on subsequent calls within refresh interval", async () => {
+    let now = 1000000;
+    const mockFetch = vi.fn().mockResolvedValue({
+      ok: true,
+      json: async () => ({
+        data: [{ id: "anthropic.claude-sonnet-4-6", object: "model" }],
+      }),
+    });
+
+    // First call — hits the network
+    const first = await discoverMantleModels({
+      region: "us-east-1",
+      bearerToken: "test-token",
+      fetchFn: mockFetch as unknown as typeof fetch,
+      now: () => now,
+    });
+    expect(first).toHaveLength(1);
+    expect(mockFetch).toHaveBeenCalledTimes(1);
+
+    // Second call within refresh interval — uses cache
+    now += 60_000; // 1 minute later
+    const second = await discoverMantleModels({
+      region: "us-east-1",
+      bearerToken: "test-token",
+      fetchFn: mockFetch as unknown as typeof fetch,
+      now: () => now,
+    });
+    expect(second).toHaveLength(1);
+    expect(mockFetch).toHaveBeenCalledTimes(1); // No additional fetch
+
+    // Third call after refresh interval — re-fetches
+    now += 3600_000; // 1 hour later
+    const third = await discoverMantleModels({
+      region: "us-east-1",
+      bearerToken: "test-token",
+      fetchFn: mockFetch as unknown as typeof fetch,
+      now: () => now,
+    });
+    expect(third).toHaveLength(1);
+    expect(mockFetch).toHaveBeenCalledTimes(2); // Re-fetched
+  });
+
+  it("returns stale cache on fetch failure", async () => {
+    let now = 1000000;
+    const mockFetch = vi
+      .fn()
+      .mockResolvedValueOnce({
+        ok: true,
+        json: async () => ({
+          data: [{ id: "anthropic.claude-sonnet-4-6", object: "model" }],
+        }),
+      })
+      .mockRejectedValueOnce(new Error("ECONNREFUSED"));
+
+    // First call — succeeds
+    await discoverMantleModels({
+      region: "us-east-1",
+      bearerToken: "test-token",
+      fetchFn: mockFetch as unknown as typeof fetch,
+      now: () => now,
+    });
+
+    // Second call after expiry — fails but returns stale cache
+    now += 7200_000;
+    const stale = await discoverMantleModels({
+      region: "us-east-1",
+      bearerToken: "test-token",
+      fetchFn: mockFetch as unknown as typeof fetch,
+      now: () => now,
+    });
+    expect(stale).toHaveLength(1);
+    expect(stale[0]?.id).toBe("anthropic.claude-sonnet-4-6");
+  });
+
+  // ---------------------------------------------------------------------------
+  // Implicit provider resolution
+  // ---------------------------------------------------------------------------
+
+  it("resolves implicit provider when bearer token is set", async () => {
+    const mockFetch = vi.fn().mockResolvedValue({
+      ok: true,
+      json: async () => ({
+        data: [{ id: "anthropic.claude-sonnet-4-6", object: "model" }],
+      }),
+    });
+
+    const provider = await resolveImplicitMantleProvider({
+      env: {
+        AWS_BEARER_TOKEN_BEDROCK: "my-token", // pragma: allowlist secret
+        AWS_REGION: "us-east-1",
+      } as NodeJS.ProcessEnv,
+      fetchFn: mockFetch as unknown as typeof fetch,
+    });
+
+    expect(provider?.baseUrl).toBe("https://bedrock-mantle.us-east-1.api.aws/v1");
+    expect(provider?.api).toBe("openai-completions");
+    expect(provider?.auth).toBe("api-key");
+    expect(provider?.apiKey).toBe("env:AWS_BEARER_TOKEN_BEDROCK");
+    expect(provider?.models).toHaveLength(3);
+    const opus = provider?.models?.find((model) => model.id === "anthropic.claude-opus-4-7");
+    expect(opus?.api).toBe("anthropic-messages");
+    expect(opus?.reasoning).toBe(false);
+    expect(opus).not.toHaveProperty("baseUrl");
+    const mythos = provider?.models?.find(
+      (model) => model.id === "anthropic.claude-mythos-preview",
+    );
+    expect(mythos).toMatchObject({
+      api: "anthropic-messages",
+      reasoning: true,
+      params: { canonicalModelId: "claude-mythos-preview" },
+      contextWindow: 1_000_000,
+      maxTokens: 128_000,
+    });
+  });
+
+  it("returns null when no auth is available", async () => {
+    const tokenProviderFactory = vi.fn(() => {
+      throw new Error("no credentials");
+    });
+
+    const provider = await resolveImplicitMantleProvider({
+      env: {} as NodeJS.ProcessEnv,
+      tokenProviderFactory,
+    });
+
+    expect(provider).toBeNull();
+  });
+
+  it("uses a generated IAM token when no explicit token is set", async () => {
+    const tokenProvider = vi.fn(async () => "bedrock-api-key-iam"); // pragma: allowlist secret
+    const tokenProviderFactory = createTokenProviderFactory(tokenProvider);
+    const mockFetch = vi.fn().mockResolvedValue({
+      ok: true,
+      json: async () => ({
+        data: [{ id: "openai.gpt-oss-120b", object: "model" }],
+      }),
+    });
+
+    const provider = await resolveImplicitMantleProvider({
+      env: {
+        AWS_PROFILE: "default",
+        AWS_REGION: "us-east-1",
+      } as NodeJS.ProcessEnv,
+      fetchFn: mockFetch as unknown as typeof fetch,
+      tokenProviderFactory,
+    });
+
+    expect(provider?.apiKey).toBe(MANTLE_IAM_TOKEN_MARKER);
+    expect(tokenProvider).toHaveBeenCalledTimes(1);
+    expect(stringArgAt(mockFetch, 0, 0)).toBe("https://bedrock-mantle.us-east-1.api.aws/v1/models");
+    expect(recordField(objectArgAt(mockFetch, 0, 1).headers, "headers").Authorization).toBe(
+      "Bearer bedrock-api-key-iam",
+    );
+  });
+
+  it("resolves Mantle runtime auth from the cached IAM token marker", async () => {
+    const tokenProvider = vi.fn(async () => "bedrock-api-key-runtime"); // pragma: allowlist secret
+    const tokenProviderFactory = createTokenProviderFactory(tokenProvider);
+
+    await generateBearerTokenFromIam({
+      region: "us-east-1",
+      now: () => 1000,
+      tokenProviderFactory,
+    });
+
+    const resolved = await resolveMantleRuntimeBearerToken({
+      apiKey: MANTLE_IAM_TOKEN_MARKER,
+      env: {
+        AWS_REGION: "us-east-1",
+      } as NodeJS.ProcessEnv,
+      now: () => 2000,
+      tokenProviderFactory,
+    });
+    expect(resolved?.apiKey).toBe("bedrock-api-key-runtime");
+    expect(resolved?.expiresAt).toBe(1000 + 7200_000);
+    expect(tokenProvider).toHaveBeenCalledTimes(1);
+  });
+
+  it("generates a fresh Mantle runtime IAM token when the cache is cold", async () => {
+    const tokenProvider = vi.fn(async () => "bedrock-api-key-fresh"); // pragma: allowlist secret
+    const tokenProviderFactory = createTokenProviderFactory(tokenProvider);
+
+    const resolved = await resolveMantleRuntimeBearerToken({
+      apiKey: MANTLE_IAM_TOKEN_MARKER,
+      env: {
+        AWS_REGION: "us-east-1",
+      } as NodeJS.ProcessEnv,
+      now: () => 5000,
+      tokenProviderFactory,
+    });
+    expect(resolved?.apiKey).toBe("bedrock-api-key-fresh");
+    expect(resolved?.expiresAt).toBe(5000 + 7200_000);
+    expect(tokenProvider).toHaveBeenCalledTimes(1);
+  });
+
+  it("omits Mantle runtime IAM token expiry when the process clock is invalid", async () => {
+    const tokenProvider = vi.fn(async () => "bedrock-api-key-invalid-clock"); // pragma: allowlist secret
+    const tokenProviderFactory = createTokenProviderFactory(tokenProvider);
+
+    const resolved = await resolveMantleRuntimeBearerToken({
+      apiKey: MANTLE_IAM_TOKEN_MARKER,
+      env: {
+        AWS_REGION: "us-east-1",
+      } as NodeJS.ProcessEnv,
+      now: () => Number.NaN,
+      tokenProviderFactory,
+    });
+    expect(resolved).toEqual({
+      apiKey: "bedrock-api-key-invalid-clock",
+    });
+    expect(tokenProvider).toHaveBeenCalledTimes(1);
+  });
+
+  it("returns null for unsupported regions", async () => {
+    const provider = await resolveImplicitMantleProvider({
+      env: {
+        AWS_BEARER_TOKEN_BEDROCK: "my-token", // pragma: allowlist secret
+        AWS_REGION: "af-south-1",
+      } as NodeJS.ProcessEnv,
+    });
+
+    expect(provider).toBeNull();
+  });
+
+  it("defaults to us-east-1 when no region is set", async () => {
+    const mockFetch = vi.fn().mockResolvedValue({
+      ok: true,
+      json: async () => ({ data: [{ id: "openai.gpt-oss-120b", object: "model" }] }),
+    });
+
+    const provider = await resolveImplicitMantleProvider({
+      env: {
+        AWS_BEARER_TOKEN_BEDROCK: "my-token", // pragma: allowlist secret
+      } as NodeJS.ProcessEnv,
+      fetchFn: mockFetch as unknown as typeof fetch,
+    });
+
+    expect(provider?.baseUrl).toBe("https://bedrock-mantle.us-east-1.api.aws/v1");
+    expect(stringArgAt(mockFetch, 0, 0)).toBe("https://bedrock-mantle.us-east-1.api.aws/v1/models");
+    objectArgAt(mockFetch, 0, 1);
+  });
+
+  // ---------------------------------------------------------------------------
+  // Provider merging
+  // ---------------------------------------------------------------------------
+
+  it("merges implicit models when existing provider has empty models", () => {
+    const result = mergeImplicitMantleProvider({
+      existing: {
+        baseUrl: "https://custom.example.com/v1",
+        models: [],
+      },
+      implicit: {
+        baseUrl: "https://bedrock-mantle.us-east-1.api.aws/v1",
+        api: "openai-completions",
+        auth: "api-key",
+        apiKey: "env:AWS_BEARER_TOKEN_BEDROCK",
+        models: [
+          {
+            id: "openai.gpt-oss-120b",
+            name: "GPT-OSS 120B",
+            reasoning: true,
+            input: ["text"],
+            cost: { input: 0, output: 0, cacheRead: 0, cacheWrite: 0 },
+            contextWindow: 32000,
+            maxTokens: 4096,
+          },
+        ],
+      },
+    });
+
+    expect(result.baseUrl).toBe("https://custom.example.com/v1");
+    expect(result.models?.map((m) => m.id)).toEqual(["openai.gpt-oss-120b"]);
+  });
+
+  it("preserves existing models over implicit ones", () => {
+    const result = mergeImplicitMantleProvider({
+      existing: {
+        baseUrl: "https://bedrock-mantle.us-east-1.api.aws/v1",
+        models: [
+          {
+            id: "custom-model",
+            name: "My Custom Model",
+            reasoning: false,
+            input: ["text"],
+            cost: { input: 0, output: 0, cacheRead: 0, cacheWrite: 0 },
+            contextWindow: 64000,
+            maxTokens: 8192,
+          },
+        ],
+      },
+      implicit: {
+        baseUrl: "https://bedrock-mantle.us-east-1.api.aws/v1",
+        api: "openai-completions",
+        auth: "api-key",
+        models: [
+          {
+            id: "openai.gpt-oss-120b",
+            name: "GPT-OSS 120B",
+            reasoning: true,
+            input: ["text"],
+            cost: { input: 0, output: 0, cacheRead: 0, cacheWrite: 0 },
+            contextWindow: 32000,
+            maxTokens: 4096,
+          },
+        ],
+      },
+    });
+
+    expect(result.models?.map((m) => m.id)).toEqual(["custom-model"]);
+  });
+});