@langwatch/mcp-server 2.1.0 → 3.0.0

package/src/__tests__/all-tools.integration.test.ts

@@ -1,1872 +0,0 @@
-import { createServer, type Server } from "http";
-import { afterAll, beforeAll, describe, expect, it, vi, beforeEach } from "vitest";
-import { initConfig } from "../config.js";
-import {
-  fetchDocumentation,
-  resolveDocumentationUrl,
-} from "../documentation-fetch.js";
-import {
-  deleteAgent,
-  getAgent,
-  updateAgent,
-} from "../langwatch-api-agents.js";
-
-// --- Canned responses for every API endpoint ---
-
-const CANNED_TRACES_SEARCH = {
-  traces: [
-    {
-      trace_id: "trace-001",
-      formatted_trace:
-        "Root [server] 1200ms\n  LLM Call [llm] 500ms\n    Input: Hello, how are you?\n    Output: I am fine, thank you!",
-      input: { value: "Hello, how are you?" },
-      output: { value: "I am fine, thank you!" },
-      timestamps: { started_at: 1700000000000 },
-      metadata: { user_id: "user-42", thread_id: "thread-1" },
-      evaluations: [
-        {
-          evaluator_id: "eval-1",
-          name: "Faithfulness",
-          score: 0.95,
-          passed: true,
-        },
-      ],
-    },
-  ],
-  pagination: { totalHits: 1 },
-};
-
-const CANNED_TRACES_SEARCH_WITH_SCROLL = {
-  traces: [
-    {
-      trace_id: "trace-page-1",
-      input: { value: "Page 1 input" },
-      output: { value: "Page 1 output" },
-    },
-  ],
-  pagination: { totalHits: 50, scrollId: "scroll-token-abc" },
-};
-
-const CANNED_TRACES_EMPTY = {
-  traces: [],
-  pagination: { totalHits: 0 },
-};
-
-const CANNED_TRACE_DETAIL = {
-  trace_id: "trace-001",
-  formatted_trace:
-    "Root [server] 1200ms\n  LLM Call [llm] 500ms\n    Input: Hello\n    Output: Hi there",
-  timestamps: {
-    started_at: 1700000000000,
-    updated_at: 1700000001000,
-    inserted_at: 1700000001000,
-  },
-  metadata: {
-    user_id: "user-42",
-    thread_id: "thread-1",
-    customer_id: "cust-100",
-    labels: ["production"],
-  },
-  evaluations: [
-    {
-      evaluator_id: "eval-1",
-      name: "Faithfulness",
-      score: 0.95,
-      passed: true,
-    },
-  ],
-};
-
-const CANNED_ANALYTICS = {
-  currentPeriod: [
-    { date: "2024-01-01", "0__trace_id_cardinality": 42 },
-    { date: "2024-01-02", "0__trace_id_cardinality": 58 },
-  ],
-  previousPeriod: [],
-};
-
-const CANNED_PROMPTS_LIST = [
-  {
-    id: "p1",
-    handle: "greeting-bot",
-    name: "Greeting Bot",
-    latestVersionNumber: 3,
-  },
-  {
-    id: "p2",
-    handle: "qa-assistant",
-    name: "QA Assistant",
-    latestVersionNumber: 1,
-  },
-];
-
-const CANNED_PROMPT_DETAIL = {
-  id: "p1",
-  handle: "greeting-bot",
-  name: "Greeting Bot",
-  version: 3,
-  versionId: "ver_p1v3",
-  commitMessage: "Updated tone",
-  model: "openai/gpt-4o",
-  messages: [{ role: "system", content: "You are a friendly bot." }],
-  parameters: {},
-  tags: [{ name: "latest", versionId: "ver_p1v3" }],
-};
-
-const CANNED_PROMPT_CREATED = {
-  id: "p-new",
-  handle: "new-prompt",
-  name: "New Prompt",
-  latestVersionNumber: 1,
-};
-
-const CANNED_PROMPT_UPDATED = {
-  id: "p1",
-  handle: "greeting-bot",
-  latestVersionNumber: 4,
-};
-
-const CANNED_SCENARIOS_LIST = [
-  {
-    id: "scen_abc123",
-    name: "Login Flow Happy Path",
-    situation: "User attempts to log in with valid credentials",
-    criteria: [
-      "Responds with a welcome message",
-      "Includes user name in greeting",
-    ],
-    labels: ["auth", "happy-path"],
-  },
-  {
-    id: "scen_def456",
-    name: "Password Reset",
-    situation: "User requests a password reset link",
-    criteria: ["Sends reset email"],
-    labels: ["auth"],
-  },
-];
-
-const CANNED_SCENARIO_DETAIL = {
-  id: "scen_abc123",
-  name: "Login Flow Happy Path",
-  situation: "User attempts to log in with valid credentials",
-  criteria: [
-    "Responds with a welcome message",
-    "Includes user name in greeting",
-  ],
-  labels: ["auth", "happy-path"],
-};
-
-const CANNED_SCENARIO_CREATED = {
-  id: "scen_new789",
-  name: "New Scenario",
-  situation: "User does something",
-  criteria: ["Agent responds correctly"],
-  labels: ["test"],
-};
-
-const CANNED_SCENARIO_UPDATED = {
-  id: "scen_abc123",
-  name: "Login Flow - Updated",
-  situation: "User logs in with correct email and pass",
-  criteria: [
-    "Responds with welcome message",
-    "Sets session cookie",
-    "Redirects to dashboard",
-  ],
-  labels: ["auth", "happy-path"],
-};
-
-const CANNED_SCENARIO_ARCHIVED = {
-  id: "scen_abc123",
-  archived: true,
-};
-
-const CANNED_EVALUATORS_LIST = [
-  {
-    id: "evaluator_abc123",
-    projectId: "proj_1",
-    name: "Toxicity Check",
-    slug: "toxicity-check",
-    type: "evaluator",
-    config: { evaluatorType: "openai/moderation" },
-    workflowId: null,
-    copiedFromEvaluatorId: null,
-    createdAt: "2024-01-01T00:00:00.000Z",
-    updatedAt: "2024-01-01T00:00:00.000Z",
-    fields: [{ identifier: "input", type: "str" }],
-    outputFields: [{ identifier: "passed", type: "bool" }],
-  },
-  {
-    id: "evaluator_def456",
-    projectId: "proj_1",
-    name: "Exact Match",
-    slug: "exact-match",
-    type: "evaluator",
-    config: { evaluatorType: "langevals/exact_match" },
-    workflowId: null,
-    copiedFromEvaluatorId: null,
-    createdAt: "2024-01-02T00:00:00.000Z",
-    updatedAt: "2024-01-02T00:00:00.000Z",
-    fields: [
-      { identifier: "output", type: "str" },
-      { identifier: "expected_output", type: "str" },
-    ],
-    outputFields: [{ identifier: "passed", type: "bool" }],
-  },
-];
-
-const CANNED_EVALUATOR_DETAIL = {
-  id: "evaluator_abc123",
-  projectId: "proj_1",
-  name: "Toxicity Check",
-  slug: "toxicity-check",
-  type: "evaluator",
-  config: {
-    evaluatorType: "openai/moderation",
-    settings: { model: "text-moderation-stable" },
-  },
-  workflowId: null,
-  copiedFromEvaluatorId: null,
-  createdAt: "2024-01-01T00:00:00.000Z",
-  updatedAt: "2024-01-01T00:00:00.000Z",
-  fields: [
-    { identifier: "input", type: "str" },
-    { identifier: "output", type: "str", optional: true },
-  ],
-  outputFields: [
-    { identifier: "passed", type: "bool" },
-    { identifier: "score", type: "float" },
-  ],
-};
-
-const CANNED_EVALUATOR_CREATED = {
-  id: "evaluator_new123",
-  projectId: "proj_1",
-  name: "My LLM Judge",
-  slug: "my-llm-judge",
-  type: "evaluator",
-  config: { evaluatorType: "langevals/llm_boolean" },
-  workflowId: null,
-  copiedFromEvaluatorId: null,
-  createdAt: "2024-01-01T00:00:00.000Z",
-  updatedAt: "2024-01-01T00:00:00.000Z",
-  fields: [{ identifier: "input", type: "str" }],
-  outputFields: [{ identifier: "passed", type: "bool" }],
-};
-
-const CANNED_EVALUATOR_UPDATED = {
-  id: "evaluator_abc123",
-  projectId: "proj_1",
-  name: "Updated Toxicity",
-  slug: "toxicity-check",
-  type: "evaluator",
-  config: { evaluatorType: "openai/moderation" },
-  workflowId: null,
-  copiedFromEvaluatorId: null,
-  createdAt: "2024-01-01T00:00:00.000Z",
-  updatedAt: "2024-01-02T00:00:00.000Z",
-  fields: [],
-  outputFields: [],
-};
-
-const CANNED_MODEL_PROVIDERS_LIST = {
-  openai: {
-    provider: "openai",
-    enabled: true,
-    customKeys: { OPENAI_API_KEY: "HAS_KEY" },
-    models: ["gpt-4o", "gpt-4o-mini"],
-    embeddingsModels: ["text-embedding-3-small"],
-    deploymentMapping: null,
-    extraHeaders: [],
-  },
-  anthropic: {
-    provider: "anthropic",
-    enabled: false,
-    customKeys: null,
-    models: ["claude-sonnet-4-5-20250929"],
-    embeddingsModels: null,
-    deploymentMapping: null,
-    extraHeaders: [],
-  },
-};
-
-const CANNED_AGENTS_LIST = {
-  data: [
-    { id: "agent_abc", name: "Test Agent", type: "http", config: { url: "http://example.com" }, createdAt: "2024-01-01T00:00:00Z", updatedAt: "2024-01-01T00:00:00Z" },
-  ],
-  pagination: { page: 1, limit: 50, total: 1, totalPages: 1 },
-};
-
-const CANNED_AGENT_DETAIL = {
-  id: "agent_abc", name: "Test Agent", type: "http", config: { url: "http://example.com" }, createdAt: "2024-01-01T00:00:00Z", updatedAt: "2024-01-01T00:00:00Z",
-};
-
-const CANNED_RUN_PLAN = {
-  id: "plan_abc", name: "Regression Plan", slug: "regression-plan", scope: { mode: "labels", labels: ["auth"] }, scenarioIds: ["scen_abc123"], targets: [{ type: "http", referenceId: "agent_abc" }], repeatCount: 2, simulatorModel: "openai/gpt-5-mini", judgeModel: null, labels: [], archivedAt: null, createdAt: "2024-01-01T00:00:00Z", updatedAt: "2024-01-01T00:00:00Z", platformUrl: "https://app.langwatch.ai/proj/agent-testing/results/regression-plan",
-};
-
-const CANNED_RUN_PLANS_LIST = [CANNED_RUN_PLAN];
-
-const CANNED_RUN_PLAN_RUN = {
-  scheduled: true, batchRunId: "batch_123", setId: "set_456", jobCount: 2, skippedArchived: { scenarios: [], targets: [] }, items: [{ scenarioRunId: "run_1", scenarioId: "scen_abc123", target: { type: "http", referenceId: "agent_abc" }, name: "Test" }], runPlanId: "plan_abc", planName: "Regression Plan", created: true, platformUrl: "https://app.langwatch.ai/proj/agent-testing/results/regression-plan",
-};
-
-const CANNED_RUN_PLAN_RERUN = { ...CANNED_RUN_PLAN_RUN, created: false };
-
-const CANNED_TEST_SUITE = {
-  id: "suite_abc", name: "Checkout", slug: "checkout", scenarioIds: ["scen_abc123"], scenarioCount: 1, archivedAt: null, createdAt: "2024-01-01T00:00:00Z", updatedAt: "2024-01-01T00:00:00Z", platformUrl: "https://app.langwatch.ai/proj/agent-testing/suites/checkout",
-};
-
-const CANNED_TEST_SUITES_LIST = [CANNED_TEST_SUITE];
-
-const CANNED_TEST_SUITE_DETAIL = {
-  ...CANNED_TEST_SUITE,
-  scenarios: [{ id: "scen_abc123", name: "Login Flow Happy Path" }],
-};
-
-const CANNED_TEST_SUITE_CREATED = {
-  ...CANNED_TEST_SUITE, id: "suite_new", name: "New Suite", slug: "new-suite", scenarioIds: [], scenarioCount: 0,
-};
-
-const CANNED_TEST_SUITE_RENAMED = { ...CANNED_TEST_SUITE, name: "Checkout v2", slug: "checkout-v2" };
-
-const CANNED_SIMULATION_RUNS = {
-  runs: [{ scenarioRunId: "run_abc", scenarioId: "scen_abc123", batchRunId: "batch_xyz", name: "Login Flow", status: "SUCCESS", durationInMs: 5200, totalCost: 0.0042, timestamp: 1700000000000, updatedAt: 1700000001000 }],
-  hasMore: false,
-};
-
-const CANNED_SIMULATION_RUN_DETAIL = {
-  scenarioRunId: "run_abc", scenarioId: "scen_abc123", batchRunId: "batch_xyz", name: "Login Flow", status: "SUCCESS", durationInMs: 5200, totalCost: 0.0042, results: { verdict: "passed", reasoning: "All criteria met", metCriteria: ["Greets user"], unmetCriteria: [], error: null }, messages: [{ role: "user", content: "Hello" }, { role: "assistant", content: "Hi there!" }], timestamp: 1700000000000, updatedAt: 1700000001000,
-};
-
-const CANNED_DASHBOARDS_LIST = {
-  data: [{ id: "dash_abc", name: "Main Dashboard", order: 0, graphCount: 3, createdAt: "2024-01-01T00:00:00Z", updatedAt: "2024-01-01T00:00:00Z" }],
-};
-
-const CANNED_WORKFLOWS_LIST = [
-  { id: "wf_abc", name: "Test Workflow", icon: null, description: "A workflow", isEvaluator: false, isComponent: false, createdAt: "2024-01-01T00:00:00Z", updatedAt: "2024-01-01T00:00:00Z" },
-];
-
-const CANNED_MODEL_PROVIDER_SET = {
-  openai: {
-    provider: "openai",
-    enabled: true,
-    customKeys: { OPENAI_API_KEY: "HAS_KEY" },
-    models: ["gpt-4o"],
-    embeddingsModels: null,
-    deploymentMapping: null,
-    extraHeaders: [],
-  },
-};
-
-// --- Mock HTTP Server that handles ALL MCP API endpoints ---
-
-/** Track last request for each route so tests can assert on request body/params. */
-const lastRequests: Record<string, { method: string; url: string; body: string }> = {};
-
-/**
- * Mutable prompt detail so the mock is stateful: a PUT updates it and the
- * subsequent confirmation GET reflects the new version. Reset per test.
- */
-let promptDetailState: typeof CANNED_PROMPT_DETAIL = { ...CANNED_PROMPT_DETAIL };
-beforeEach(() => {
-  promptDetailState = { ...CANNED_PROMPT_DETAIL };
-});
-
-function createMockServer(): Server {
-  return createServer((req, res) => {
-    const authToken = req.headers["x-auth-token"];
-    if (authToken !== "test-integration-key") {
-      res.writeHead(401, { "Content-Type": "application/json" });
-      res.end(JSON.stringify({ message: "Invalid auth token." }));
-      return;
-    }
-
-    let body = "";
-    req.on("data", (chunk: string) => (body += chunk));
-    req.on("end", () => {
-      const url = req.url ?? "";
-      const method = req.method ?? "GET";
-      res.setHeader("Content-Type", "application/json");
-
-      // Store last request for assertions
-      const routeKey = `${method} ${url.split("?")[0]}`;
-      lastRequests[routeKey] = { method, url, body };
-
-      // --- Trace endpoints ---
-      if (url === "/api/traces/search" && method === "POST") {
-        const parsed = JSON.parse(body);
-        // Return empty results when a special query is used
-        if (parsed.query === "__empty__") {
-          res.writeHead(200);
-          res.end(JSON.stringify(CANNED_TRACES_EMPTY));
-        } else if (parsed.pageSize === 5) {
-          res.writeHead(200);
-          res.end(JSON.stringify(CANNED_TRACES_SEARCH_WITH_SCROLL));
-        } else {
-          res.writeHead(200);
-          res.end(JSON.stringify(CANNED_TRACES_SEARCH));
-        }
-      } else if (
-        url.match(/^\/api\/traces\/trace-nonexistent(\?|$)/) &&
-        method === "GET"
-      ) {
-        res.writeHead(404);
-        res.end(JSON.stringify({ message: "Trace not found" }));
-      } else if (
-        url.match(/^\/api\/traces\/[^/]+(\?|$)/) &&
-        method === "GET"
-      ) {
-        res.writeHead(200);
-        res.end(JSON.stringify(CANNED_TRACE_DETAIL));
-      }
-      // --- Analytics endpoint ---
-      else if (url === "/api/analytics/timeseries" && method === "POST") {
-        res.writeHead(200);
-        res.end(JSON.stringify(CANNED_ANALYTICS));
-      }
-      // --- Prompt endpoints ---
-      else if (url === "/api/prompts" && method === "GET") {
-        res.writeHead(200);
-        res.end(JSON.stringify(CANNED_PROMPTS_LIST));
-      } else if (url === "/api/prompts" && method === "POST") {
-        res.writeHead(200);
-        res.end(JSON.stringify(CANNED_PROMPT_CREATED));
-      } else if (
-        url.match(/^\/api\/prompts\/[^/]+$/) &&
-        method === "GET"
-      ) {
-        res.writeHead(200);
-        res.end(JSON.stringify(promptDetailState));
-      } else if (
-        url.match(/^\/api\/prompts\/[^/]+$/) &&
-        method === "PUT"
-      ) {
-        const parsed = JSON.parse(body) as {
-          commitMessage?: string;
-          model?: string;
-          messages?: Array<{ role: string; content: string }>;
-          tags?: string[];
-        };
-        const newVersionId = "ver_p1v4";
-        promptDetailState = {
-          ...promptDetailState,
-          version: 4,
-          versionId: newVersionId,
-          commitMessage: parsed.commitMessage ?? promptDetailState.commitMessage,
-          model: parsed.model ?? promptDetailState.model,
-          messages: parsed.messages ?? promptDetailState.messages,
-          tags:
-            parsed.tags === undefined
-              ? promptDetailState.tags.map((tag) =>
-                  tag.name === "latest"
-                    ? { name: "latest", versionId: newVersionId }
-                    : tag,
-                )
-              : [
-                  { name: "latest", versionId: newVersionId },
-                  ...parsed.tags
-                    .filter((name) => name !== "latest")
-                    .map((name) => ({ name, versionId: newVersionId })),
-                ],
-        };
-        res.writeHead(200);
-        res.end(JSON.stringify(CANNED_PROMPT_UPDATED));
-      }
-      // --- Scenario endpoints ---
-      else if (url === "/api/scenarios" && method === "GET") {
-        res.writeHead(200);
-        res.end(JSON.stringify(CANNED_SCENARIOS_LIST));
-      } else if (url === "/api/scenarios" && method === "POST") {
-        res.writeHead(200);
-        res.end(JSON.stringify(CANNED_SCENARIO_CREATED));
-      } else if (
-        url.match(/^\/api\/scenarios\/scen_nonexistent(\?|$)/) &&
-        method === "GET"
-      ) {
-        res.writeHead(404);
-        res.end(JSON.stringify({ message: "Scenario not found" }));
-      } else if (
-        url.match(/^\/api\/scenarios\/[^/]+$/) &&
-        method === "GET"
-      ) {
-        res.writeHead(200);
-        res.end(JSON.stringify(CANNED_SCENARIO_DETAIL));
-      } else if (
-        url.match(/^\/api\/scenarios\/[^/]+$/) &&
-        method === "PUT"
-      ) {
-        res.writeHead(200);
-        res.end(JSON.stringify(CANNED_SCENARIO_UPDATED));
-      } else if (
-        url.match(/^\/api\/scenarios\/[^/]+$/) &&
-        method === "DELETE"
-      ) {
-        res.writeHead(200);
-        res.end(JSON.stringify(CANNED_SCENARIO_ARCHIVED));
-      }
-      // --- Evaluator endpoints ---
-      else if (url === "/api/evaluators" && method === "GET") {
-        res.writeHead(200);
-        res.end(JSON.stringify(CANNED_EVALUATORS_LIST));
-      } else if (url === "/api/evaluators" && method === "POST") {
-        res.writeHead(200);
-        res.end(JSON.stringify(CANNED_EVALUATOR_CREATED));
-      } else if (
-        url.match(/^\/api\/evaluators\/evaluator_nonexistent(\?|$)/) &&
-        method === "GET"
-      ) {
-        res.writeHead(404);
-        res.end(JSON.stringify({ message: "Evaluator not found" }));
-      } else if (
-        url.match(/^\/api\/evaluators\/[^/]+$/) &&
-        method === "GET"
-      ) {
-        res.writeHead(200);
-        res.end(JSON.stringify(CANNED_EVALUATOR_DETAIL));
-      } else if (
-        url.match(/^\/api\/evaluators\/[^/]+$/) &&
-        method === "PUT"
-      ) {
-        res.writeHead(200);
-        res.end(JSON.stringify(CANNED_EVALUATOR_UPDATED));
-      }
-      // --- Model Provider endpoints ---
-      else if (url === "/api/model-providers" && method === "GET") {
-        res.writeHead(200);
-        res.end(JSON.stringify(CANNED_MODEL_PROVIDERS_LIST));
-      } else if (
-        url.match(/^\/api\/model-providers\/[^/]+$/) &&
-        method === "PUT"
-      ) {
-        res.writeHead(200);
-        res.end(JSON.stringify(CANNED_MODEL_PROVIDER_SET));
-      }
-      // --- Agent endpoints ---
-      else if (url === "/api/v1/agents" && method === "GET") {
-        res.writeHead(200);
-        res.end(JSON.stringify(CANNED_AGENTS_LIST));
-      } else if (url === "/api/v1/agents" && method === "POST") {
-        res.writeHead(201);
-        res.end(JSON.stringify(CANNED_AGENT_DETAIL));
-      } else if (url?.match(/^\/api\/v1\/agents\/[^/]+$/) && method === "GET") {
-        res.writeHead(200);
-        res.end(JSON.stringify(CANNED_AGENT_DETAIL));
-      } else if (url?.match(/^\/api\/v1\/agents\/[^/]+$/) && method === "PATCH") {
-        res.writeHead(200);
-        res.end(JSON.stringify(CANNED_AGENT_DETAIL));
-      } else if (url?.match(/^\/api\/v1\/agents\/[^/]+$/) && method === "DELETE") {
-        res.writeHead(200);
-        res.end(JSON.stringify({ id: "agent_abc", name: "Test Agent" }));
-      }
-      // --- Run plan endpoints ---
-      else if (url === "/api/v1/run-plans/run" && method === "POST") {
-        res.writeHead(200);
-        res.end(JSON.stringify(CANNED_RUN_PLAN_RUN));
-      } else if (url?.match(/^\/api\/v1\/run-plans(\?.*)?$/) && method === "GET") {
-        res.writeHead(200);
-        res.end(JSON.stringify(CANNED_RUN_PLANS_LIST));
-      } else if (url?.match(/^\/api\/v1\/run-plans\/[^/]+\/run$/) && method === "POST") {
-        res.writeHead(200);
-        res.end(JSON.stringify(CANNED_RUN_PLAN_RERUN));
-      } else if (url?.match(/^\/api\/v1\/run-plans\/[^/]+$/) && method === "GET") {
-        res.writeHead(200);
-        res.end(JSON.stringify(CANNED_RUN_PLAN));
-      } else if (url?.match(/^\/api\/v1\/run-plans\/[^/]+$/) && method === "DELETE") {
-        res.writeHead(200);
-        res.end(JSON.stringify({ id: "plan_abc", archived: true }));
-      }
-      // --- Test suite endpoints ---
-      else if (url === "/api/v1/test-suites" && method === "GET") {
-        res.writeHead(200);
-        res.end(JSON.stringify(CANNED_TEST_SUITES_LIST));
-      } else if (url === "/api/v1/test-suites" && method === "POST") {
-        res.writeHead(201);
-        res.end(JSON.stringify(CANNED_TEST_SUITE_CREATED));
-      } else if (url?.match(/^\/api\/v1\/test-suites\/[^/]+\/run$/) && method === "POST") {
-        res.writeHead(200);
-        res.end(JSON.stringify(CANNED_RUN_PLAN_RUN));
-      } else if (url?.match(/^\/api\/v1\/test-suites\/[^/]+$/) && method === "GET") {
-        res.writeHead(200);
-        res.end(JSON.stringify(CANNED_TEST_SUITE_DETAIL));
-      } else if (url?.match(/^\/api\/v1\/test-suites\/[^/]+$/) && method === "PATCH") {
-        res.writeHead(200);
-        res.end(JSON.stringify(CANNED_TEST_SUITE_RENAMED));
-      } else if (url?.match(/^\/api\/v1\/test-suites\/[^/]+$/) && method === "DELETE") {
-        res.writeHead(200);
-        res.end(JSON.stringify({ id: "suite_abc", archived: true }));
-      }
-      // --- Simulation Run endpoints ---
-      else if (url?.match(/^\/api\/simulation-runs\/[^/]+$/) && method === "GET") {
-        res.writeHead(200);
-        res.end(JSON.stringify(CANNED_SIMULATION_RUN_DETAIL));
-      } else if (url?.match(/^\/api\/simulation-runs/) && method === "GET") {
-        res.writeHead(200);
-        res.end(JSON.stringify(CANNED_SIMULATION_RUNS));
-      }
-      // --- Dashboard endpoints ---
-      else if (url === "/api/dashboards" && method === "GET") {
-        res.writeHead(200);
-        res.end(JSON.stringify(CANNED_DASHBOARDS_LIST));
-      } else if (url === "/api/dashboards" && method === "POST") {
-        res.writeHead(201);
-        res.end(JSON.stringify({ id: "dash_new", name: "New Dashboard" }));
-      } else if (url?.match(/^\/api\/dashboards\/[^/]+$/) && method === "GET") {
-        res.writeHead(200);
-        res.end(JSON.stringify(CANNED_DASHBOARDS_LIST.data[0]));
-      } else if (url?.match(/^\/api\/dashboards\/[^/]+$/) && method === "DELETE") {
-        res.writeHead(200);
-        res.end(JSON.stringify({ id: "dash_abc", name: "Main Dashboard" }));
-      }
-      // --- Workflow endpoints ---
-      else if (url === "/api/workflows" && method === "GET") {
-        res.writeHead(200);
-        res.end(JSON.stringify(CANNED_WORKFLOWS_LIST));
-      } else if (url?.match(/^\/api\/workflows\/[^/]+\/run$/) && method === "POST") {
-        res.writeHead(200);
-        res.end(JSON.stringify({ output: "workflow result" }));
-      } else if (url?.match(/^\/api\/workflows\/[^/]+$/) && method === "GET") {
-        res.writeHead(200);
-        res.end(JSON.stringify(CANNED_WORKFLOWS_LIST[0]));
-      } else if (url?.match(/^\/api\/workflows\/[^/]+$/) && method === "DELETE") {
-        res.writeHead(200);
-        res.end(JSON.stringify({ id: "wf_abc", archived: true }));
-      }
-      // --- Monitor endpoints ---
-      else if (url === "/api/monitors" && method === "GET") {
-        res.writeHead(200);
-        res.end(JSON.stringify([
-          { id: "mon_abc", name: "Toxicity Check", slug: "toxicity-check-x1y2z", checkType: "ragas/toxicity", enabled: true, executionMode: "ON_MESSAGE", sample: 1.0, level: "trace", evaluatorId: null, preconditions: [], parameters: {}, mappings: {}, threadIdleTimeout: null, createdAt: "2026-01-01T00:00:00Z", updatedAt: "2026-01-01T00:00:00Z" },
-        ]));
-      } else if (url === "/api/monitors" && method === "POST") {
-        res.writeHead(201);
-        res.end(JSON.stringify({ id: "mon_new", name: "New Monitor", slug: "new-monitor-abc12", checkType: "ragas/toxicity", enabled: true, executionMode: "ON_MESSAGE", sample: 1.0, level: "trace", evaluatorId: null, preconditions: [], parameters: {}, mappings: {}, threadIdleTimeout: null, createdAt: "2026-01-01T00:00:00Z", updatedAt: "2026-01-01T00:00:00Z" }));
-      } else if (url?.match(/^\/api\/monitors\/[^/]+\/toggle$/) && method === "POST") {
-        const monitorId = url?.split("/")[3];
-        res.writeHead(200);
-        res.end(JSON.stringify({ id: monitorId, enabled: true }));
-      } else if (url?.match(/^\/api\/monitors\/[^/]+$/) && method === "GET") {
-        res.writeHead(200);
-        res.end(JSON.stringify({ id: "mon_abc", name: "Toxicity Check", slug: "toxicity-check-x1y2z", checkType: "ragas/toxicity", enabled: true, executionMode: "ON_MESSAGE", sample: 1.0, level: "trace", evaluatorId: null, preconditions: [], parameters: {}, mappings: {}, threadIdleTimeout: null, createdAt: "2026-01-01T00:00:00Z", updatedAt: "2026-01-01T00:00:00Z" }));
-      } else if (url?.match(/^\/api\/monitors\/[^/]+$/) && method === "PATCH") {
-        res.writeHead(200);
-        res.end(JSON.stringify({ id: "mon_abc", name: "Updated Monitor", slug: "toxicity-check-x1y2z", checkType: "ragas/toxicity", enabled: false, executionMode: "ON_MESSAGE", sample: 0.5, level: "trace", evaluatorId: null, preconditions: [], parameters: {}, mappings: {}, threadIdleTimeout: null, createdAt: "2026-01-01T00:00:00Z", updatedAt: "2026-01-02T00:00:00Z" }));
-      } else if (url?.match(/^\/api\/monitors\/[^/]+$/) && method === "DELETE") {
-        const monitorId = url?.split("/").pop();
-        res.writeHead(200);
-        res.end(JSON.stringify({ id: monitorId, deleted: true }));
-      }
-      // --- Secret endpoints ---
-      else if (url === "/api/secrets" && method === "GET") {
-        res.writeHead(200);
-        res.end(JSON.stringify([
-          { id: "secret_abc", projectId: "proj_123", name: "MY_API_KEY", createdAt: "2026-01-01T00:00:00Z", updatedAt: "2026-01-01T00:00:00Z" },
-          { id: "secret_def", projectId: "proj_123", name: "DB_PASSWORD", createdAt: "2026-01-02T00:00:00Z", updatedAt: "2026-01-02T00:00:00Z" },
-        ]));
-      } else if (url === "/api/secrets" && method === "POST") {
-        res.writeHead(201);
-        res.end(JSON.stringify({ id: "secret_new", projectId: "proj_123", name: "NEW_SECRET", createdAt: "2026-01-03T00:00:00Z", updatedAt: "2026-01-03T00:00:00Z" }));
-      } else if (url?.match(/^\/api\/secrets\/[^/]+$/) && method === "PUT") {
-        res.writeHead(200);
-        res.end(JSON.stringify({ id: "secret_abc", projectId: "proj_123", name: "MY_API_KEY", createdAt: "2026-01-01T00:00:00Z", updatedAt: "2026-01-03T00:00:00Z" }));
-      } else if (url?.match(/^\/api\/secrets\/[^/]+$/) && method === "DELETE") {
-        const secretId = url?.split("/").pop();
-        res.writeHead(200);
-        res.end(JSON.stringify({ id: secretId, deleted: true }));
-      }
-      // --- Fallback ---
-      else {
-        res.writeHead(404);
-        res.end(
-          JSON.stringify({ message: `Not found: ${method} ${url}` }),
-        );
-      }
-    });
-  });
-}
-
-// --- Integration Tests ---
-// These verify that every MCP tool handler correctly communicates with the REST API
-// through a real HTTP server. Formatting/digest assertions ensure the full chain works.
-
-describe("All MCP tools integration", () => {
-  let server: Server;
-  let port: number;
-
-  beforeAll(async () => {
-    server = createMockServer();
-    await new Promise<void>((resolve) => {
-      server.listen(0, () => {
-        const addr = server.address();
-        port = typeof addr === "object" && addr ? addr.port : 0;
-        initConfig({
-          apiKey: "test-integration-key",
-          endpoint: `http://localhost:${port}`,
-        });
-        resolve();
-      });
-    });
-  });
-
-  afterAll(async () => {
-    await new Promise<void>((resolve) => server.close(() => resolve()));
-  });
-
-  // =====================
-  // 1. fetch_langwatch_docs
-  // =====================
-  describe("fetch_langwatch_docs", () => {
-    it("resolves the trusted docs index", () => {
-      expect(resolveDocumentationUrl("langwatch").toString()).toBe(
-        "https://langwatch.ai/docs/llms.txt",
-      );
-    });
-
-    it("fetches trusted docs without following redirects", async () => {
-      const mockFetch = vi.fn().mockResolvedValue(
-        new Response("# Integration Guide"),
-      );
-
-      const text = await fetchDocumentation(
-        "langwatch",
-        "integration",
-        mockFetch,
-      );
-
-      expect(text).toContain("Integration Guide");
-      expect(mockFetch).toHaveBeenCalledWith(
-        new URL("https://langwatch.ai/docs/integration.md"),
-        {
-          redirect: "error",
-          signal: expect.any(AbortSignal),
-        },
-      );
-    });
-  });
-
-  // =====================
-  // 2. fetch_scenario_docs
-  // =====================
-  describe("fetch_scenario_docs", () => {
-    it("resolves the trusted Scenario docs index", () => {
-      expect(resolveDocumentationUrl("scenario").toString()).toBe(
-        "https://langwatch.ai/scenario/llms.txt",
-      );
-    });
-
-    it("normalizes a relative Scenario docs path", () => {
-      expect(resolveDocumentationUrl("scenario", "setup").toString()).toBe(
-        "https://langwatch.ai/scenario/setup.md",
-      );
-    });
-  });
-
-  // =====================
-  // 3. discover_schema
-  // =====================
-  describe("discover_schema", () => {
-    describe("when category is filters", () => {
-      /** @scenario Agent discovers available filter fields */
-      it("returns filter field documentation", async () => {
-        const { formatSchema } = await import(
-          "../tools/discover-schema.js"
-        );
-        const result = formatSchema("filters");
-
-        expect(result).toContain("## Available Filter Fields");
-        expect(result).toContain("filters");
-      });
-    });
-
-    describe("when category is metrics", () => {
-      /** @scenario Agent discovers available metrics with allowed aggregations */
-      it("returns metric documentation", async () => {
-        const { formatSchema } = await import(
-          "../tools/discover-schema.js"
-        );
-        const result = formatSchema("metrics");
-
-        expect(result).toContain("## Available Metrics");
-      });
-    });
-
-    describe("when category is aggregations", () => {
-      it("returns aggregation types", async () => {
-        const { formatSchema } = await import(
-          "../tools/discover-schema.js"
-        );
-        const result = formatSchema("aggregations");
-
-        expect(result).toContain("## Available Aggregation Types");
-        expect(result).toContain("cardinality");
-        expect(result).toContain("avg");
-        expect(result).toContain("sum");
-      });
-    });
-
-    describe("when category is groups", () => {
-      /** @scenario Agent discovers available group-by options */
-      it("returns group-by options", async () => {
-        const { formatSchema } = await import(
-          "../tools/discover-schema.js"
-        );
-        const result = formatSchema("groups");
-
-        expect(result).toContain("## Available Group-By Options");
-      });
-    });
-
-    describe("when category is scenarios", () => {
-      it("returns scenario schema documentation", async () => {
-        const { formatScenarioSchema } = await import(
-          "../tools/discover-scenario-schema.js"
-        );
-        const result = formatScenarioSchema();
-
-        expect(result).toContain("# Scenario Schema");
-        expect(result).toContain("**name** (required)");
-        expect(result).toContain("**situation** (required)");
-      });
-    });
-
-    describe("when category is evaluators", () => {
-      it("returns evaluator type overview", async () => {
-        const { formatEvaluatorSchema } = await import(
-          "../tools/discover-evaluator-schema.js"
-        );
-        const result = formatEvaluatorSchema();
-
-        expect(result).toContain("# Available Evaluator Types");
-      });
-    });
-
-    describe("when category is evaluators with specific type", () => {
-      it("returns detailed evaluator schema", async () => {
-        const { formatEvaluatorSchema } = await import(
-          "../tools/discover-evaluator-schema.js"
-        );
-        const result = formatEvaluatorSchema("langevals/llm_boolean");
-
-        expect(result).toContain("langevals/llm_boolean");
-        expect(result).toContain("## Settings");
-      });
-    });
-
-    describe("when evaluator type is unknown", () => {
-      it("returns an error message", async () => {
-        const { formatEvaluatorSchema } = await import(
-          "../tools/discover-evaluator-schema.js"
-        );
-        const result = formatEvaluatorSchema("nonexistent/type");
-
-        expect(result).toContain('Unknown evaluator type');
-      });
-    });
-
-    describe("when category is all", () => {
-      /** @scenario Agent discovers all schema information at once */
-      it("returns all schema categories", async () => {
-        const { formatSchema } = await import(
-          "../tools/discover-schema.js"
-        );
-        const result = formatSchema("all");
-
-        expect(result).toContain("## Available Filter Fields");
-        expect(result).toContain("## Available Metrics");
-        expect(result).toContain("## Available Aggregation Types");
-        expect(result).toContain("## Available Group-By Options");
-      });
-    });
-  });
-
-  // =====================
-  // 4. search_traces
-  // =====================
-  describe("search_traces", () => {
-    describe("when traces are found", () => {
-      /** @scenario Agent searches traces with a text query */
-      it("returns formatted trace digests", async () => {
-        const { handleSearchTraces } = await import(
-          "../tools/search-traces.js"
-        );
-        const result = await handleSearchTraces({
-          startDate: "24h",
-          endDate: "now",
-        });
-
-        expect(result).toContain("trace-001");
-        expect(result).toContain("LLM Call [llm] 500ms");
-        expect(result).toContain("1 trace");
-      });
-    });
-
-    describe("when no traces match", () => {
-      it("returns a no-results message", async () => {
-        const { handleSearchTraces } = await import(
-          "../tools/search-traces.js"
-        );
-        const result = await handleSearchTraces({
-          query: "__empty__",
-        });
-
-        expect(result).toBe("No traces found matching your query.");
-      });
-    });
-
-    describe("when pagination token is present", () => {
-      /** @scenario Agent paginates through trace results */
-      it("includes scroll ID for next page", async () => {
-        const { handleSearchTraces } = await import(
-          "../tools/search-traces.js"
-        );
-        const result = await handleSearchTraces({
-          pageSize: 5,
-        });
-
-        expect(result).toContain("scroll-token-abc");
-      });
-    });
-
-    describe("when format is json", () => {
-      it("returns parseable JSON", async () => {
-        const { handleSearchTraces } = await import(
-          "../tools/search-traces.js"
-        );
-        const result = await handleSearchTraces({
-          format: "json",
-        });
-
-        const parsed = JSON.parse(result);
-        expect(parsed.traces).toBeDefined();
-        expect(parsed.traces.length).toBeGreaterThan(0);
-        expect(parsed.traces[0].trace_id).toBe("trace-001");
-      });
-    });
-
-    describe("when filters are applied", () => {
-      /** @scenario Agent searches traces filtered by user_id */
-      it("passes filters to the API", async () => {
-        const { handleSearchTraces } = await import(
-          "../tools/search-traces.js"
-        );
-        await handleSearchTraces({
-          filters: { "metadata.user_id": ["user-42"] },
-        });
-
-        const req = lastRequests["POST /api/traces/search"];
-        expect(req).toBeDefined();
-        const parsed = JSON.parse(req!.body);
-        expect(parsed.filters).toEqual({
-          "metadata.user_id": ["user-42"],
-        });
-      });
-    });
-
-    describe("given a trace has evaluation results", () => {
-      /** @scenario Agent searches traces and sees evaluation results without a follow-up call */
-      it("includes evaluation status in the digest without a follow-up call", async () => {
-        const { handleSearchTraces } = await import(
-          "../tools/search-traces.js"
-        );
-        const result = await handleSearchTraces({
-          startDate: "24h",
-          endDate: "now",
-        });
-
-        expect(result).toContain("- **Faithfulness**: PASSED (score: 0.95)");
-      });
-    });
-  });
-
-  // =====================
-  // 5. get_trace
-  // =====================
-  describe("get_trace", () => {
-    describe("when trace exists", () => {
-      /** @scenario Agent gets a single trace by ID in AI-readable format */
-      it("returns formatted trace with metadata and evaluations", async () => {
-        const { handleGetTrace } = await import(
-          "../tools/get-trace.js"
-        );
-        const result = await handleGetTrace({ traceId: "trace-001" });
-
-        expect(result).toContain("# Trace: trace-001");
-        expect(result).toContain("LLM Call [llm] 500ms");
-        expect(result).toContain("Faithfulness");
-        expect(result).toContain("PASSED");
-        expect(result).toContain("**User**: user-42");
-        expect(result).toContain("**Thread**: thread-1");
-        expect(result).toContain("**Customer**: cust-100");
-      });
-    });
-
-    describe("when trace does not exist", () => {
-      /** @scenario Agent gets a trace that does not exist */
-      it("propagates the 404 error", async () => {
-        const { handleGetTrace } = await import(
-          "../tools/get-trace.js"
-        );
-
-        await expect(
-          handleGetTrace({ traceId: "trace-nonexistent" }),
-        ).rejects.toThrow("404");
-      });
-    });
-
-    describe("when format is json", () => {
-      it("returns parseable JSON with full trace data", async () => {
-        const { handleGetTrace } = await import(
-          "../tools/get-trace.js"
-        );
-        const result = await handleGetTrace({
-          traceId: "trace-001",
-          format: "json",
-        });
-
-        const parsed = JSON.parse(result);
-        expect(parsed.trace_id).toBe("trace-001");
-        expect(parsed.evaluations).toBeDefined();
-        expect(parsed.metadata).toBeDefined();
-      });
-    });
-  });
-
-  // =====================
-  // 6. get_analytics
-  // =====================
-  describe("get_analytics", () => {
-    describe("when data is available", () => {
-      it("returns formatted analytics with markdown table", async () => {
-        const { handleGetAnalytics } = await import(
-          "../tools/get-analytics.js"
-        );
-        const result = await handleGetAnalytics({
-          metric: "metadata.trace_id",
-          aggregation: "cardinality",
-          startDate: "7d",
-        });
-
-        expect(result).toContain("42");
-        expect(result).toContain("58");
-        expect(result).toContain("| Date | Value |");
-      });
-    });
-
-    describe("when metric and aggregation are specified", () => {
-      it("passes them through to the API", async () => {
-        const { handleGetAnalytics } = await import(
-          "../tools/get-analytics.js"
-        );
-        await handleGetAnalytics({
-          metric: "performance.total_cost",
-          aggregation: "sum",
-        });
-
-        const req = lastRequests["POST /api/analytics/timeseries"];
-        expect(req).toBeDefined();
-        const parsed = JSON.parse(req!.body);
-        expect(parsed.series[0].metric).toBe("performance.total_cost");
-        expect(parsed.series[0].aggregation).toBe("sum");
-      });
-    });
-  });
-
-  // =====================
-  // 7. platform_create_prompt
-  // =====================
-  describe("platform_create_prompt", () => {
-    describe("when valid data is provided", () => {
-      it("returns success confirmation with prompt details", async () => {
-        const { handleCreatePrompt } = await import(
-          "../tools/create-prompt.js"
-        );
-        const result = await handleCreatePrompt({
-          name: "New Prompt",
-          messages: [{ role: "system", content: "You are helpful." }],
-          model: "openai/gpt-4o",
-        });
-
-        expect(result).toContain("created successfully");
-        expect(result).toContain("p-new");
-        expect(result).toContain("**Name**: New Prompt");
-        expect(result).toContain("**Model**: openai/gpt-4o");
-      });
-    });
-  });
-
-  // =====================
-  // 8. platform_list_prompts
-  // =====================
-  describe("platform_list_prompts", () => {
-    describe("when prompts exist", () => {
-      it("returns formatted prompt list", async () => {
-        const { handleListPrompts } = await import(
-          "../tools/list-prompts.js"
-        );
-        const result = await handleListPrompts();
-
-        expect(result).toContain("greeting-bot");
-        expect(result).toContain("Greeting Bot");
-        expect(result).toContain("qa-assistant");
-        expect(result).toContain("QA Assistant");
-        expect(result).toContain("# Prompts (2 total)");
-      });
-    });
-  });
-
-  // =====================
-  // 9. platform_get_prompt
-  // =====================
-  describe("platform_get_prompt", () => {
-    describe("when prompt exists", () => {
-      it("returns formatted prompt details with messages and deployments", async () => {
-        const { handleGetPrompt } = await import(
-          "../tools/get-prompt.js"
-        );
-        const result = await handleGetPrompt({
-          idOrHandle: "greeting-bot",
-        });
-
-        expect(result).toContain("# Prompt: Greeting Bot");
-        expect(result).toContain("gpt-4o");
-        expect(result).toContain("You are a friendly bot.");
-        expect(result).toContain("v3");
-        expect(result).toContain("## Deployments");
-        expect(result).toContain("Updated tone");
-      });
-    });
-  });
-
-  // =====================
-  // 10. platform_update_prompt
-  // =====================
-  describe("platform_update_prompt", () => {
-    describe("when updating a prompt", () => {
-      it("returns success message", async () => {
-        const { handleUpdatePrompt } = await import(
-          "../tools/update-prompt.js"
-        );
-        const result = await handleUpdatePrompt({
-          idOrHandle: "greeting-bot",
-          model: "openai/gpt-4o-mini",
-          commitMessage: "Switch to mini",
-        });
-
-        expect(result).toContain("updated successfully");
-        expect(result).toContain("Switch to mini");
-        expect(result).toContain("**Version**: v4");
-        expect(result).toContain("**Version ID**: ver_p1v4");
-      });
-    });
-
-    describe("when an update supplies only messages and a commit message", () => {
-      /** @scenario Carrying forward prior fields when an update supplies only messages and a commit message */
-      it("does not send fields that would wipe prior prompt configuration", async () => {
-        const { handleUpdatePrompt } = await import(
-          "../tools/update-prompt.js"
-        );
-        await handleUpdatePrompt({
-          idOrHandle: "greeting-bot",
-          messages: [{ role: "system", content: "You are a helpful bot." }],
-          commitMessage: "Only messages changed",
-        });
-
-        const req = lastRequests["PUT /api/prompts/greeting-bot"];
-        expect(req).toBeDefined();
-        const parsed = JSON.parse(req!.body);
-        expect(parsed).not.toHaveProperty("parameters");
-        expect(parsed).not.toHaveProperty("inputs");
-        expect(parsed).not.toHaveProperty("outputs");
-        expect(parsed).not.toHaveProperty("temperature");
-        expect(parsed).not.toHaveProperty("responseFormat");
-      });
-    });
-
-    describe("when an update omits tags", () => {
-      /** @scenario Tag-to-version mapping stays unchanged when an update omits tags */
-      it("does not send a tags field in the update request", async () => {
-        const { handleUpdatePrompt } = await import(
-          "../tools/update-prompt.js"
-        );
-        await handleUpdatePrompt({
-          idOrHandle: "greeting-bot",
-          model: "openai/gpt-4o",
-          commitMessage: "No tag change requested",
-        });
-
-        const req = lastRequests["PUT /api/prompts/greeting-bot"];
-        expect(req).toBeDefined();
-        const parsed = JSON.parse(req!.body);
-        expect(parsed).not.toHaveProperty("tags");
-      });
-    });
-
-    describe("when an update passes tags explicitly", () => {
-      /** @scenario Passing tags explicitly moves the tag to the new version */
-      it("sends the requested tags in the update request", async () => {
-        const { handleUpdatePrompt } = await import(
-          "../tools/update-prompt.js"
-        );
-        await handleUpdatePrompt({
-          idOrHandle: "greeting-bot",
-          model: "openai/gpt-4o",
-          commitMessage: "Deploy new version to production",
-          tags: ["production"],
-        });
-
-        const req = lastRequests["PUT /api/prompts/greeting-bot"];
-        expect(req).toBeDefined();
-        const parsed = JSON.parse(req!.body);
-        expect(parsed.tags).toEqual(["production"]);
-      });
-    });
-  });
-
-  // =====================
-  // 11. platform_create_scenario
-  // =====================
-  describe("platform_create_scenario", () => {
-    describe("when valid data is provided", () => {
-      it("returns confirmation with new scenario ID", async () => {
-        const { handleCreateScenario } = await import(
-          "../tools/create-scenario.js"
-        );
-        const result = await handleCreateScenario({
-          name: "New Scenario",
-          situation: "User does something",
-          criteria: ["Agent responds correctly"],
-          labels: ["test"],
-        });
-
-        expect(result).toContain("created successfully");
-        expect(result).toContain("scen_new789");
-      });
-    });
-  });
-
-  // =====================
-  // 12. platform_list_scenarios
-  // =====================
-  describe("platform_list_scenarios", () => {
-    describe("when scenarios exist", () => {
-      it("returns formatted scenario list", async () => {
-        const { handleListScenarios } = await import(
-          "../tools/list-scenarios.js"
-        );
-        const result = await handleListScenarios({});
-
-        expect(result).toContain("# Scenarios (2 total)");
-        expect(result).toContain("Login Flow Happy Path");
-        expect(result).toContain("Password Reset");
-      });
-    });
-
-    describe("when format is json", () => {
-      it("returns parseable JSON matching API response", async () => {
-        const { handleListScenarios } = await import(
-          "../tools/list-scenarios.js"
-        );
-        const result = await handleListScenarios({ format: "json" });
-
-        expect(JSON.parse(result)).toEqual(CANNED_SCENARIOS_LIST);
-      });
-    });
-  });
-
-  // =====================
-  // 13. platform_get_scenario
-  // =====================
-  describe("platform_get_scenario", () => {
-    describe("when the scenario exists", () => {
-      it("returns formatted scenario details", async () => {
-        const { handleGetScenario } = await import(
-          "../tools/get-scenario.js"
-        );
-        const result = await handleGetScenario({
-          scenarioId: "scen_abc123",
-        });
-
-        expect(result).toContain("# Scenario: Login Flow Happy Path");
-        expect(result).toContain("User attempts to log in");
-        expect(result).toContain("Responds with a welcome message");
-      });
-    });
-
-    describe("when the scenario does not exist", () => {
-      it("propagates the 404 error", async () => {
-        const { handleGetScenario } = await import(
-          "../tools/get-scenario.js"
-        );
-
-        await expect(
-          handleGetScenario({ scenarioId: "scen_nonexistent" }),
-        ).rejects.toThrow("404");
-      });
-    });
-
-    describe("when format is json", () => {
-      it("returns parseable JSON", async () => {
-        const { handleGetScenario } = await import(
-          "../tools/get-scenario.js"
-        );
-        const result = await handleGetScenario({
-          scenarioId: "scen_abc123",
-          format: "json",
-        });
-
-        expect(JSON.parse(result)).toEqual(CANNED_SCENARIO_DETAIL);
-      });
-    });
-  });
-
-  // =====================
-  // 14. platform_update_scenario
-  // =====================
-  describe("platform_update_scenario", () => {
-    describe("when the scenario exists", () => {
-      it("returns update confirmation with updated details", async () => {
-        const { handleUpdateScenario } = await import(
-          "../tools/update-scenario.js"
-        );
-        const result = await handleUpdateScenario({
-          scenarioId: "scen_abc123",
-          name: "Login Flow - Updated",
-        });
-
-        expect(result).toContain("updated successfully");
-        expect(result).toContain("scen_abc123");
-      });
-    });
-  });
-
-  // =====================
-  // 15. platform_archive_scenario
-  // =====================
-  describe("platform_archive_scenario", () => {
-    describe("when the scenario exists", () => {
-      it("returns confirmation that scenario was archived", async () => {
-        const { handleArchiveScenario } = await import(
-          "../tools/archive-scenario.js"
-        );
-        const result = await handleArchiveScenario({
-          scenarioId: "scen_abc123",
-        });
-
-        expect(result).toContain("archived successfully");
-        expect(result).toContain("scen_abc123");
-        expect(result).toContain("archived");
-      });
-    });
-  });
-
-  // =====================
-  // 16. platform_create_evaluator
-  // =====================
-  describe("platform_create_evaluator", () => {
-    describe("when valid data is provided", () => {
-      it("returns success confirmation with evaluator details", async () => {
-        const { handleCreateEvaluator } = await import(
-          "../tools/create-evaluator.js"
-        );
-        const result = await handleCreateEvaluator({
-          name: "My LLM Judge",
-          config: { evaluatorType: "langevals/llm_boolean" },
-        });
-
-        expect(result).toContain("Evaluator created successfully!");
-        expect(result).toContain("evaluator_new123");
-        expect(result).toContain("my-llm-judge");
-        expect(result).toContain("langevals/llm_boolean");
-      });
-    });
-  });
-
-  // =====================
-  // 17. platform_list_evaluators
-  // =====================
-  describe("platform_list_evaluators", () => {
-    describe("when evaluators exist", () => {
-      it("returns formatted evaluator list", async () => {
-        const { handleListEvaluators } = await import(
-          "../tools/list-evaluators.js"
-        );
-        const result = await handleListEvaluators();
-
-        expect(result).toContain("# Evaluators (2 total)");
-        expect(result).toContain("Toxicity Check");
-        expect(result).toContain("toxicity-check");
-        expect(result).toContain("openai/moderation");
-        expect(result).toContain("Exact Match");
-        expect(result).toContain("exact-match");
-      });
-    });
-  });
-
-  // =====================
-  // 18. platform_get_evaluator
-  // =====================
-  describe("platform_get_evaluator", () => {
-    describe("when the evaluator exists", () => {
-      it("returns formatted evaluator details with config and fields", async () => {
-        const { handleGetEvaluator } = await import(
-          "../tools/get-evaluator.js"
-        );
-        const result = await handleGetEvaluator({
-          idOrSlug: "evaluator_abc123",
-        });
-
-        expect(result).toContain("# Evaluator: Toxicity Check");
-        expect(result).toContain("openai/moderation");
-        expect(result).toContain("text-moderation-stable");
-        expect(result).toContain("## Input Fields");
-        expect(result).toContain("**input** (str)");
-        expect(result).toContain("## Output Fields");
-        expect(result).toContain("**passed** (bool)");
-        expect(result).toContain("**score** (float)");
-      });
-    });
-
-    describe("when the evaluator does not exist", () => {
-      it("propagates the 404 error", async () => {
-        const { handleGetEvaluator } = await import(
-          "../tools/get-evaluator.js"
-        );
-
-        await expect(
-          handleGetEvaluator({ idOrSlug: "evaluator_nonexistent" }),
-        ).rejects.toThrow("404");
-      });
-    });
-  });
-
-  // =====================
-  // 19. platform_update_evaluator
-  // =====================
-  describe("platform_update_evaluator", () => {
-    describe("when the evaluator exists", () => {
-      it("returns update confirmation", async () => {
-        const { handleUpdateEvaluator } = await import(
-          "../tools/update-evaluator.js"
-        );
-        const result = await handleUpdateEvaluator({
-          evaluatorId: "evaluator_abc123",
-          name: "Updated Toxicity",
-        });
-
-        expect(result).toContain("Evaluator updated successfully!");
-        expect(result).toContain("evaluator_abc123");
-        expect(result).toContain("Updated Toxicity");
-      });
-    });
-  });
-
-  // =====================
-  // 20. platform_set_model_provider
-  // =====================
-  describe("platform_set_model_provider", () => {
-    describe("when setting a provider with API key", () => {
-      it("returns success confirmation with provider details", async () => {
-        const { handleSetModelProvider } = await import(
-          "../tools/set-model-provider.js"
-        );
-        const result = await handleSetModelProvider({
-          provider: "openai",
-          enabled: true,
-          customKeys: { OPENAI_API_KEY: "sk-test123" },
-        });
-
-        expect(result).toContain("Model provider updated successfully!");
-        expect(result).toContain("**Provider**: openai");
-        expect(result).toContain("**Status**: enabled");
-        expect(result).toContain("OPENAI_API_KEY: set");
-      });
-    });
-
-    describe("when setting a default model", () => {
-      it("shows the normalized model name with provider prefix", async () => {
-        const { handleSetModelProvider } = await import(
-          "../tools/set-model-provider.js"
-        );
-        const result = await handleSetModelProvider({
-          provider: "openai",
-          enabled: true,
-          defaultModel: "gpt-4o",
-        });
-
-        expect(result).toContain("**Default Model**: openai/gpt-4o");
-      });
-    });
-  });
-
-  // =====================
-  // 21. platform_list_model_providers
-  // =====================
-  describe("platform_list_model_providers", () => {
-    describe("when providers exist", () => {
-      it("returns formatted provider list with status and key info", async () => {
-        const { handleListModelProviders } = await import(
-          "../tools/list-model-providers.js"
-        );
-        const result = await handleListModelProviders();
-
-        expect(result).toContain("# Model Providers (2 total)");
-        expect(result).toContain("## openai");
-        expect(result).toContain("**Status**: enabled");
-        expect(result).toContain("## anthropic");
-        expect(result).toContain("**Status**: disabled");
-        expect(result).toContain("OPENAI_API_KEY: set");
-        expect(result).toContain("2 available");
-      });
-    });
-  });
-
-  // =====================
-  // Cross-cutting: authentication
-  // =====================
-  describe("when API key is invalid", () => {
-    afterAll(() => {
-      initConfig({
-        apiKey: "test-integration-key",
-        endpoint: `http://localhost:${port}`,
-      });
-    });
-
-    it("throws an error with 401 status for trace search", async () => {
-      initConfig({
-        apiKey: "bad-key",
-        endpoint: `http://localhost:${port}`,
-      });
-
-      const { handleSearchTraces } = await import(
-        "../tools/search-traces.js"
-      );
-      await expect(
-        handleSearchTraces({ startDate: "24h" }),
-      ).rejects.toThrow("401");
-    });
-
-    it("throws an error with 401 status for evaluator list", async () => {
-      initConfig({
-        apiKey: "bad-key",
-        endpoint: `http://localhost:${port}`,
-      });
-
-      const { handleListEvaluators } = await import(
-        "../tools/list-evaluators.js"
-      );
-      await expect(handleListEvaluators()).rejects.toThrow("401");
-    });
-
-    it("throws an error with 401 status for model providers list", async () => {
-      initConfig({
-        apiKey: "bad-key",
-        endpoint: `http://localhost:${port}`,
-      });
-
-      const { handleListModelProviders } = await import(
-        "../tools/list-model-providers.js"
-      );
-      await expect(handleListModelProviders()).rejects.toThrow("401");
-    });
-  });
-
-  // =====================
-  // Run Plan Tools
-  // =====================
-  describe("platform_run_plan", () => {
-    it("runs a configuration and reports the plan it created", async () => {
-      const { handleRunPlan } = await import("../tools/run-plan.js");
-      const result = await handleRunPlan({
-        name: "Regression Plan",
-        scope: { mode: "labels", labels: ["auth"] },
-        targets: [{ type: "http", referenceId: "agent_abc" }],
-        repeatCount: 2,
-        note: "nightly regression",
-      });
-
-      expect(result).toContain("Regression Plan");
-      expect(result).toContain("created and started");
-      expect(result).toContain("batch_123");
-      expect(result).toContain("**Jobs**: 2");
-    });
-  });
-
-  describe("platform_list_run_plans", () => {
-    it("returns formatted run plan list", async () => {
-      const { handleListRunPlans } = await import("../tools/list-run-plans.js");
-      const result = await handleListRunPlans({});
-
-      expect(result).toContain("Run Plans (1 total)");
-      expect(result).toContain("Regression Plan");
-    });
-  });
-
-  describe("platform_get_run_plan", () => {
-    it("returns the plan configuration", async () => {
-      const { handleGetRunPlan } = await import("../tools/get-run-plan.js");
-      const result = await handleGetRunPlan({ id: "plan_abc" });
-
-      expect(result).toContain("Regression Plan");
-      expect(result).toContain("plan_abc");
-      expect(result).toContain("labels: auth");
-    });
-  });
-
-  describe("platform_rerun_run_plan", () => {
-    it("runs the stored configuration again and reports it joined the plan", async () => {
-      const { handleRerunRunPlan } = await import(
-        "../tools/rerun-run-plan.js"
-      );
-      const result = await handleRerunRunPlan({ id: "plan_abc" });
-
-      expect(result).toContain("Regression Plan");
-      expect(result).toContain("configuration of this run");
-      expect(result).toContain("batch_123");
-    });
-  });
-
-  describe("platform_archive_run_plan", () => {
-    it("archives the run plan", async () => {
-      const { handleArchiveRunPlan } = await import(
-        "../tools/archive-run-plan.js"
-      );
-      const result = await handleArchiveRunPlan({ id: "plan_abc" });
-
-      expect(result).toContain("archived");
-      expect(result).toContain("plan_abc");
-    });
-  });
-
-  // =====================
-  // Test Suite Tools
-  // =====================
-  describe("platform_list_test_suites", () => {
-    it("returns formatted test suite list", async () => {
-      const { handleListTestSuites } = await import(
-        "../tools/list-test-suites.js"
-      );
-      const result = await handleListTestSuites({});
-
-      expect(result).toContain("Test Suites (1 total)");
-      expect(result).toContain("Checkout");
-    });
-  });
-
-  describe("platform_create_test_suite", () => {
-    it("creates a test suite and returns confirmation", async () => {
-      const { handleCreateTestSuite } = await import(
-        "../tools/create-test-suite.js"
-      );
-      const result = await handleCreateTestSuite({ name: "New Suite" });
-
-      expect(result).toContain("created");
-      expect(result).toContain("New Suite");
-      expect(result).toContain("suite_new");
-    });
-  });
-
-  describe("platform_get_test_suite", () => {
-    it("returns the suite with the scenarios filed in it", async () => {
-      const { handleGetTestSuite } = await import(
-        "../tools/get-test-suite.js"
-      );
-      const result = await handleGetTestSuite({ id: "suite_abc" });
-
-      expect(result).toContain("Checkout");
-      expect(result).toContain("Login Flow Happy Path");
-    });
-  });
-
-  describe("platform_rename_test_suite", () => {
-    it("renames the test suite", async () => {
-      const { handleRenameTestSuite } = await import(
-        "../tools/rename-test-suite.js"
-      );
-      const result = await handleRenameTestSuite({
-        id: "suite_abc",
-        name: "Checkout v2",
-      });
-
-      expect(result).toContain("Checkout v2");
-    });
-  });
-
-  describe("platform_archive_test_suite", () => {
-    it("archives the suite and says the filed scenarios went with it", async () => {
-      const { handleArchiveTestSuite } = await import(
-        "../tools/archive-test-suite.js"
-      );
-      const result = await handleArchiveTestSuite({ id: "suite_abc" });
-
-      expect(result).toContain("archived");
-      expect(result).toContain("scenarios filed in it");
-    });
-  });
-
-  describe("platform_run_test_suite", () => {
-    it("runs the suite against a target and reports the derived plan", async () => {
-      const { handleRunTestSuite } = await import(
-        "../tools/run-test-suite.js"
-      );
-      const result = await handleRunTestSuite({
-        id: "suite_abc",
-        targets: [{ type: "http", referenceId: "agent_abc" }],
-      });
-
-      expect(result).toContain("Regression Plan");
-      expect(result).toContain("batch_123");
-    });
-  });
-
-  // =====================
-  // Simulation Run Tools
-  // =====================
-  describe("platform_list_simulation_runs", () => {
-    it("returns formatted run list", async () => {
-      const { handleListSimulationRuns } = await import("../tools/list-simulation-runs.js");
-      const result = await handleListSimulationRuns({});
-
-      expect(result).toContain("Simulation Runs");
-      expect(result).toContain("Login Flow");
-    });
-  });
-
-  describe("platform_get_simulation_run", () => {
-    it("returns run details with conversation", async () => {
-      const { handleGetSimulationRun } = await import("../tools/get-simulation-run.js");
-      const result = await handleGetSimulationRun({ scenarioRunId: "run_abc" });
-
-      expect(result).toContain("Login Flow");
-      expect(result).toContain("passed");
-      expect(result).toContain("Hello");
-    });
-  });
-
-  // =====================
-  // Monitor Tools
-  // =====================
-  describe("platform_list_monitors", () => {
-    it("returns formatted monitor list", async () => {
-      const { listMonitors } = await import("../langwatch-api-monitors.js");
-      const monitors = await listMonitors();
-      expect(monitors).toHaveLength(1);
-      expect(monitors[0]!.name).toBe("Toxicity Check");
-      expect(monitors[0]!.checkType).toBe("ragas/toxicity");
-    });
-  });
-
-  describe("platform_get_monitor", () => {
-    it("returns monitor details", async () => {
-      const { getMonitor } = await import("../langwatch-api-monitors.js");
-      const monitor = await getMonitor("mon_abc");
-      expect(monitor.id).toBe("mon_abc");
-      expect(monitor.name).toBe("Toxicity Check");
-      expect(monitor.enabled).toBe(true);
-    });
-  });
-
-  describe("platform_create_monitor", () => {
-    it("creates a monitor and returns metadata", async () => {
-      const { createMonitor } = await import("../langwatch-api-monitors.js");
-      const monitor = await createMonitor({ name: "New Monitor", checkType: "ragas/toxicity" });
-      expect(monitor.id).toBe("mon_new");
-      expect(monitor.name).toBe("New Monitor");
-    });
-  });
-
-  describe("platform_update_monitor", () => {
-    it("updates a monitor", async () => {
-      const { updateMonitor } = await import("../langwatch-api-monitors.js");
-      const monitor = await updateMonitor({ id: "mon_abc", enabled: false });
-      expect(monitor.id).toBe("mon_abc");
-      expect(monitor.enabled).toBe(false);
-    });
-  });
-
-  describe("platform_delete_monitor", () => {
-    it("deletes a monitor", async () => {
-      const { deleteMonitor } = await import("../langwatch-api-monitors.js");
-      const result = await deleteMonitor("mon_abc");
-      expect(result.id).toBe("mon_abc");
-      expect(result.deleted).toBe(true);
-    });
-  });
-
-  // =====================
-  // Secret Tools
-  // =====================
-  describe("platform_list_secrets", () => {
-    it("returns formatted secret list", async () => {
-      const { listSecrets } = await import("../langwatch-api-secrets.js");
-      const secrets = await listSecrets();
-      expect(secrets).toHaveLength(2);
-      expect(secrets[0]!.name).toBe("MY_API_KEY");
-    });
-  });
-
-  describe("platform_create_secret", () => {
-    it("creates a secret and returns metadata", async () => {
-      const { createSecret } = await import("../langwatch-api-secrets.js");
-      const secret = await createSecret({ name: "NEW_SECRET", value: "sk-123" });
-      expect(secret.id).toBe("secret_new");
-      expect(secret.name).toBe("NEW_SECRET");
-    });
-  });
-
-  describe("platform_update_secret", () => {
-    it("updates a secret value", async () => {
-      const { updateSecret } = await import("../langwatch-api-secrets.js");
-      const secret = await updateSecret({ id: "secret_abc", value: "new-val" });
-      expect(secret.id).toBe("secret_abc");
-      expect(secret.name).toBe("MY_API_KEY");
-    });
-  });
-
-  describe("platform_delete_secret", () => {
-    it("deletes a secret", async () => {
-      const { deleteSecret } = await import("../langwatch-api-secrets.js");
-      const result = await deleteSecret("secret_abc");
-      expect(result.id).toBe("secret_abc");
-      expect(result.deleted).toBe(true);
-    });
-  });
-
-  // =====================
-  // Agent Tools
-  // =====================
-  describe("platform_get_agent", () => {
-    it("returns one agent by id", async () => {
-      const agent = await getAgent("agent_abc");
-      expect(agent.id).toBe("agent_abc");
-      expect(agent.name).toBe("Test Agent");
-    });
-  });
-
-  describe("platform_update_agent", () => {
-    it("updates an agent and returns it", async () => {
-      const agent = await updateAgent({ id: "agent_abc", name: "Test Agent" });
-      expect(agent.id).toBe("agent_abc");
-    });
-  });
-
-  describe("platform_delete_agent", () => {
-    it("deletes an agent", async () => {
-      const result = await deleteAgent("agent_abc");
-      expect(result.id).toBe("agent_abc");
-    });
-  });
-
-  // =====================
-  // Agent Run Tool
-  // =====================
-  describe("platform_run_workflow", () => {
-    it("executes a workflow and returns result", async () => {
-      const { handleRunWorkflow } = await import("../tools/run-workflow.js");
-      const result = await handleRunWorkflow({ id: "wf_abc" });
-
-      expect(result).toContain("executed successfully");
-      expect(result).toContain("workflow result");
-    });
-  });
-});