@langwatch/mcp-server 0.4.0 → 0.5.0

package/CHANGELOG.md

@@ -1,5 +1,24 @@
 # Changelog
 
+## [0.5.0](https://github.com/langwatch/langwatch/compare/mcp-server@v0.4.0...mcp-server@v0.5.0) (2026-02-20)
+
+
+### Features
+
+* add scenario management tools to MCP server ([#1705](https://github.com/langwatch/langwatch/issues/1705)) ([0376fde](https://github.com/langwatch/langwatch/commit/0376fde0abff7b110b4ec5996a399c4b2ceafde0))
+
+
+### Miscellaneous
+
+* **deps-dev:** bump @eslint/js from 9.35.0 to 9.39.2 in /mcp-server ([#1465](https://github.com/langwatch/langwatch/issues/1465)) ([fbee07d](https://github.com/langwatch/langwatch/commit/fbee07d8b964d0a059eaa32c7685c8bf667898e7))
+* **deps:** bump hono ([f7e8f05](https://github.com/langwatch/langwatch/commit/f7e8f056843958cac4504ae02f37a351457f77ee))
+* **deps:** bump hono from 4.11.9 to 4.12.0 in /mcp-server in the npm_and_yarn group across 1 directory ([#1736](https://github.com/langwatch/langwatch/issues/1736)) ([f7e8f05](https://github.com/langwatch/langwatch/commit/f7e8f056843958cac4504ae02f37a351457f77ee))
+* **deps:** bump node-pty from 1.0.0 to 1.1.0 in /mcp-server ([#1447](https://github.com/langwatch/langwatch/issues/1447)) ([12ad02c](https://github.com/langwatch/langwatch/commit/12ad02c19dcc0ba90ad32f77659816b768188a53))
+* **deps:** bump qs ([f0e9747](https://github.com/langwatch/langwatch/commit/f0e97475becd58dfa523a944fbb3fa0657dfc1dc))
+* **deps:** bump qs from 6.14.1 to 6.14.2 in /mcp-server in the npm_and_yarn group across 1 directory ([#1568](https://github.com/langwatch/langwatch/issues/1568)) ([f0e9747](https://github.com/langwatch/langwatch/commit/f0e97475becd58dfa523a944fbb3fa0657dfc1dc))
+* **deps:** bump the npm_and_yarn group across 1 directory with 8 updates ([#1519](https://github.com/langwatch/langwatch/issues/1519)) ([487e563](https://github.com/langwatch/langwatch/commit/487e5637a941fa9335ec8e951efdf38bb0a02a8c))
+* **deps:** bump the uv group across 1 directory with 7 updates ([#1516](https://github.com/langwatch/langwatch/issues/1516)) ([7f2f178](https://github.com/langwatch/langwatch/commit/7f2f178588d89a63f3b38510844b87de7b528b3b))
+
 ## [0.4.0](https://github.com/langwatch/langwatch/compare/mcp-server@v0.3.3...mcp-server@v0.4.0) (2026-02-08)
 
 

package/dist/archive-scenario-GAE4XVFM.js

@@ -0,0 +1,19 @@
+import {
+  archiveScenario
+} from "./chunk-K2YFPOSD.js";
+import "./chunk-JVWDWL3J.js";
+import "./chunk-AAQNA53E.js";
+
+// src/tools/archive-scenario.ts
+async function handleArchiveScenario(params) {
+  const result = await archiveScenario(params.scenarioId);
+  const lines = [];
+  lines.push("Scenario archived successfully!\n");
+  lines.push(`**ID**: ${result.id}`);
+  lines.push(`**Status**: ${result.archived ? "archived" : "active"}`);
+  return lines.join("\n");
+}
+export {
+  handleArchiveScenario
+};
+//# sourceMappingURL=archive-scenario-GAE4XVFM.js.map

package/dist/archive-scenario-GAE4XVFM.js.map

@@ -0,0 +1 @@
+{"version":3,"sources":["../src/tools/archive-scenario.ts"],"sourcesContent":["import { archiveScenario as apiArchiveScenario } from \"../langwatch-api-scenarios.js\";\n\n/**\n * Handles the archive_scenario MCP tool invocation.\n *\n * Archives (soft-deletes) a scenario and returns confirmation.\n */\nexport async function handleArchiveScenario(params: {\n  scenarioId: string;\n}): Promise<string> {\n  const result = await apiArchiveScenario(params.scenarioId);\n\n  const lines: string[] = [];\n  lines.push(\"Scenario archived successfully!\\n\");\n  lines.push(`**ID**: ${result.id}`);\n  lines.push(`**Status**: ${result.archived ? \"archived\" : \"active\"}`);\n\n  return lines.join(\"\\n\");\n}\n"],"mappings":";;;;;;;AAOA,eAAsB,sBAAsB,QAExB;AAClB,QAAM,SAAS,MAAM,gBAAmB,OAAO,UAAU;AAEzD,QAAM,QAAkB,CAAC;AACzB,QAAM,KAAK,mCAAmC;AAC9C,QAAM,KAAK,WAAW,OAAO,EAAE,EAAE;AACjC,QAAM,KAAK,eAAe,OAAO,WAAW,aAAa,QAAQ,EAAE;AAEnE,SAAO,MAAM,KAAK,IAAI;AACxB;","names":[]}

package/dist/{chunk-HOPTUDCZ.js → chunk-JVWDWL3J.js}

@@ -9,7 +9,7 @@ async function makeRequest(method, path, body) {
   const headers = {
     "X-Auth-Token": requireApiKey()
   };
-  if (method === "POST") {
+  if (body !== void 0) {
     headers["Content-Type"] = "application/json";
   }
   const response = await fetch(url, {
@@ -78,6 +78,7 @@ async function createPromptVersion(idOrHandle, data) {
 }
 
 export {
+  makeRequest,
   searchTraces,
   getTraceById,
   getAnalyticsTimeseries,
@@ -87,4 +88,4 @@ export {
   updatePrompt,
   createPromptVersion
 };
-//# sourceMappingURL=chunk-HOPTUDCZ.js.map
+//# sourceMappingURL=chunk-JVWDWL3J.js.map

package/dist/chunk-JVWDWL3J.js.map

@@ -0,0 +1 @@
+{"version":3,"sources":["../src/langwatch-api.ts"],"sourcesContent":["import { getConfig, requireApiKey } from \"./config.js\";\n\n// --- Response types ---\n\nexport interface TraceSearchResult {\n  trace_id: string;\n  formatted_trace?: string;\n  input?: { value: string };\n  output?: { value: string };\n  timestamps?: { started_at?: string | number };\n  metadata?: Record<string, unknown>;\n  error?: Record<string, unknown>;\n}\n\nexport interface SearchTracesResponse {\n  traces: TraceSearchResult[];\n  pagination?: {\n    totalHits?: number;\n    scrollId?: string;\n  };\n}\n\nexport interface TraceDetailResponse {\n  trace_id: string;\n  formatted_trace?: string;\n  input?: { value: string };\n  output?: { value: string };\n  timestamps?: {\n    started_at?: string | number;\n    updated_at?: string | number;\n    inserted_at?: string | number;\n  };\n  metadata?: {\n    user_id?: string;\n    thread_id?: string;\n    customer_id?: string;\n    labels?: string[];\n    [key: string]: unknown;\n  };\n  error?: Record<string, unknown>;\n  ascii_tree?: string;\n  evaluations?: Array<{\n    evaluator_id?: string;\n    name?: string;\n    score?: number;\n    passed?: boolean;\n    label?: string;\n  }>;\n  spans?: Array<{\n    span_id: string;\n    name?: string;\n    type?: string;\n    model?: string;\n    input?: { value: string };\n    output?: { value: string };\n    timestamps?: { started_at?: number; finished_at?: number };\n    metrics?: {\n      completion_time_ms?: number;\n      prompt_tokens?: number;\n      completion_tokens?: number;\n      tokens_estimated?: boolean;\n      cost?: number;\n    };\n  }>;\n}\n\nexport interface AnalyticsBucket {\n  date: string;\n  [key: string]: unknown;\n}\n\nexport interface AnalyticsTimeseriesResponse {\n  currentPeriod: AnalyticsBucket[];\n  previousPeriod: AnalyticsBucket[];\n}\n\nexport interface PromptSummary {\n  id?: string;\n  handle?: string;\n  name?: string;\n  description?: string | null;\n  latestVersionNumber?: number;\n  version?: number;\n}\n\nexport interface PromptVersion {\n  version?: number;\n  commitMessage?: string;\n  model?: string;\n  modelProvider?: string;\n  messages?: Array<{ role: string; content: string }>;\n}\n\nexport interface PromptDetailResponse extends PromptSummary {\n  versions?: PromptVersion[];\n  model?: string;\n  modelProvider?: string;\n  messages?: Array<{ role: string; content: string }>;\n  prompt?: Array<{ role: string; content: string }>;\n}\n\nexport interface PromptMutationResponse {\n  id?: string;\n  handle?: string;\n  name?: string;\n  latestVersionNumber?: number;\n}\n\n// --- HTTP client ---\n\n/**\n * Sends an HTTP request to the LangWatch API.\n *\n * Builds the full URL from the configured endpoint, adds authentication,\n * and handles JSON serialization/deserialization.\n *\n * @throws Error with status code and response body when the response is not OK\n */\nexport async function makeRequest(\n  method: \"GET\" | \"POST\" | \"PUT\" | \"PATCH\" | \"DELETE\",\n  path: string,\n  body?: unknown\n): Promise<unknown> {\n  const url = getConfig().endpoint + path;\n  const headers: Record<string, string> = {\n    \"X-Auth-Token\": requireApiKey(),\n  };\n\n  if (body !== undefined) {\n    headers[\"Content-Type\"] = \"application/json\";\n  }\n\n  const response = await fetch(url, {\n    method,\n    headers,\n    ...(body !== undefined ? { body: JSON.stringify(body) } : {}),\n  });\n\n  if (!response.ok) {\n    const responseBody = await response.text();\n    throw new Error(\n      `LangWatch API error ${response.status}: ${responseBody}`\n    );\n  }\n\n  return response.json();\n}\n\n/** Searches traces with optional filters and pagination. */\nexport async function searchTraces(params: {\n  query?: string;\n  filters?: Record<string, string[]>;\n  startDate: number;\n  endDate: number;\n  pageSize?: number;\n  pageOffset?: number;\n  scrollId?: string;\n  format?: \"digest\" | \"json\";\n}): Promise<SearchTracesResponse> {\n  const { format = \"digest\", ...rest } = params;\n  return makeRequest(\"POST\", \"/api/traces/search\", {\n    ...rest,\n    format,\n  }) as Promise<SearchTracesResponse>;\n}\n\n/** Retrieves a single trace by its ID. */\nexport async function getTraceById(\n  traceId: string,\n  format: \"digest\" | \"json\" = \"digest\"\n): Promise<TraceDetailResponse> {\n  return makeRequest(\n    \"GET\",\n    `/api/traces/${encodeURIComponent(traceId)}?format=${format}`\n  ) as Promise<TraceDetailResponse>;\n}\n\n/** Fetches analytics timeseries data for the given metrics and date range. */\nexport async function getAnalyticsTimeseries(params: {\n  series: Array<{\n    metric: string;\n    aggregation: string;\n    key?: string;\n    subkey?: string;\n  }>;\n  startDate: number;\n  endDate: number;\n  timeZone?: string;\n  groupBy?: string;\n  groupByKey?: string;\n  filters?: Record<string, string[]>;\n}): Promise<AnalyticsTimeseriesResponse> {\n  return makeRequest(\n    \"POST\",\n    \"/api/analytics/timeseries\",\n    params\n  ) as Promise<AnalyticsTimeseriesResponse>;\n}\n\n/** Lists all prompts in the project. */\nexport async function listPrompts(): Promise<PromptSummary[]> {\n  return makeRequest(\"GET\", \"/api/prompts\") as Promise<PromptSummary[]>;\n}\n\n/** Retrieves a single prompt by ID or handle. */\nexport async function getPrompt(\n  idOrHandle: string,\n  version?: number\n): Promise<PromptDetailResponse> {\n  const query = version != null ? `?version=${version}` : \"\";\n  return makeRequest(\n    \"GET\",\n    `/api/prompts/${encodeURIComponent(idOrHandle)}${query}`\n  ) as Promise<PromptDetailResponse>;\n}\n\n/** Creates a new prompt. */\nexport async function createPrompt(data: {\n  name: string;\n  handle?: string;\n  messages: Array<{ role: string; content: string }>;\n  model: string;\n  modelProvider: string;\n  description?: string;\n}): Promise<PromptMutationResponse> {\n  return makeRequest(\n    \"POST\",\n    \"/api/prompts\",\n    data\n  ) as Promise<PromptMutationResponse>;\n}\n\n/** Updates an existing prompt by ID or handle. */\nexport async function updatePrompt(\n  idOrHandle: string,\n  data: {\n    messages?: Array<{ role: string; content: string }>;\n    model?: string;\n    modelProvider?: string;\n    commitMessage?: string;\n  }\n): Promise<PromptMutationResponse> {\n  return makeRequest(\n    \"POST\",\n    `/api/prompts/${encodeURIComponent(idOrHandle)}`,\n    data\n  ) as Promise<PromptMutationResponse>;\n}\n\n/** Creates a new version of an existing prompt. */\nexport async function createPromptVersion(\n  idOrHandle: string,\n  data: {\n    messages?: Array<{ role: string; content: string }>;\n    model?: string;\n    modelProvider?: string;\n    commitMessage?: string;\n  }\n): Promise<PromptMutationResponse> {\n  return makeRequest(\n    \"POST\",\n    `/api/prompts/${encodeURIComponent(idOrHandle)}/versions`,\n    data\n  ) as Promise<PromptMutationResponse>;\n}\n\n"],"mappings":";;;;;;AAsHA,eAAsB,YACpB,QACA,MACA,MACkB;AAClB,QAAM,MAAM,UAAU,EAAE,WAAW;AACnC,QAAM,UAAkC;AAAA,IACtC,gBAAgB,cAAc;AAAA,EAChC;AAEA,MAAI,SAAS,QAAW;AACtB,YAAQ,cAAc,IAAI;AAAA,EAC5B;AAEA,QAAM,WAAW,MAAM,MAAM,KAAK;AAAA,IAChC;AAAA,IACA;AAAA,IACA,GAAI,SAAS,SAAY,EAAE,MAAM,KAAK,UAAU,IAAI,EAAE,IAAI,CAAC;AAAA,EAC7D,CAAC;AAED,MAAI,CAAC,SAAS,IAAI;AAChB,UAAM,eAAe,MAAM,SAAS,KAAK;AACzC,UAAM,IAAI;AAAA,MACR,uBAAuB,SAAS,MAAM,KAAK,YAAY;AAAA,IACzD;AAAA,EACF;AAEA,SAAO,SAAS,KAAK;AACvB;AAGA,eAAsB,aAAa,QASD;AAChC,QAAM,EAAE,SAAS,UAAU,GAAG,KAAK,IAAI;AACvC,SAAO,YAAY,QAAQ,sBAAsB;AAAA,IAC/C,GAAG;AAAA,IACH;AAAA,EACF,CAAC;AACH;AAGA,eAAsB,aACpB,SACA,SAA4B,UACE;AAC9B,SAAO;AAAA,IACL;AAAA,IACA,eAAe,mBAAmB,OAAO,CAAC,WAAW,MAAM;AAAA,EAC7D;AACF;AAGA,eAAsB,uBAAuB,QAaJ;AACvC,SAAO;AAAA,IACL;AAAA,IACA;AAAA,IACA;AAAA,EACF;AACF;AAGA,eAAsB,cAAwC;AAC5D,SAAO,YAAY,OAAO,cAAc;AAC1C;AAGA,eAAsB,UACpB,YACA,SAC+B;AAC/B,QAAM,QAAQ,WAAW,OAAO,YAAY,OAAO,KAAK;AACxD,SAAO;AAAA,IACL;AAAA,IACA,gBAAgB,mBAAmB,UAAU,CAAC,GAAG,KAAK;AAAA,EACxD;AACF;AAGA,eAAsB,aAAa,MAOC;AAClC,SAAO;AAAA,IACL;AAAA,IACA;AAAA,IACA;AAAA,EACF;AACF;AAGA,eAAsB,aACpB,YACA,MAMiC;AACjC,SAAO;AAAA,IACL;AAAA,IACA,gBAAgB,mBAAmB,UAAU,CAAC;AAAA,IAC9C;AAAA,EACF;AACF;AAGA,eAAsB,oBACpB,YACA,MAMiC;AACjC,SAAO;AAAA,IACL;AAAA,IACA,gBAAgB,mBAAmB,UAAU,CAAC;AAAA,IAC9C;AAAA,EACF;AACF;","names":[]}

package/dist/chunk-K2YFPOSD.js

@@ -0,0 +1,40 @@
+import {
+  makeRequest
+} from "./chunk-JVWDWL3J.js";
+
+// src/langwatch-api-scenarios.ts
+async function listScenarios() {
+  return makeRequest("GET", "/api/scenarios");
+}
+async function getScenario(id) {
+  return makeRequest(
+    "GET",
+    `/api/scenarios/${encodeURIComponent(id)}`
+  );
+}
+async function createScenario(data) {
+  return makeRequest("POST", "/api/scenarios", data);
+}
+async function updateScenario(params) {
+  const { id, ...data } = params;
+  return makeRequest(
+    "PUT",
+    `/api/scenarios/${encodeURIComponent(id)}`,
+    data
+  );
+}
+async function archiveScenario(id) {
+  return makeRequest(
+    "DELETE",
+    `/api/scenarios/${encodeURIComponent(id)}`
+  );
+}
+
+export {
+  listScenarios,
+  getScenario,
+  createScenario,
+  updateScenario,
+  archiveScenario
+};
+//# sourceMappingURL=chunk-K2YFPOSD.js.map

package/dist/chunk-K2YFPOSD.js.map

@@ -0,0 +1 @@
+{"version":3,"sources":["../src/langwatch-api-scenarios.ts"],"sourcesContent":["import { makeRequest } from \"./langwatch-api.js\";\n\n// --- Scenario types ---\n\nexport interface ScenarioSummary {\n  id: string;\n  name: string;\n  situation: string;\n  criteria: string[];\n  labels: string[];\n}\n\nexport interface ScenarioArchiveResponse {\n  id: string;\n  archived: boolean;\n}\n\n// --- Scenario API functions ---\n\n/** Lists all scenarios in the project. */\nexport async function listScenarios(): Promise<ScenarioSummary[]> {\n  return makeRequest(\"GET\", \"/api/scenarios\") as Promise<ScenarioSummary[]>;\n}\n\n/** Retrieves a single scenario by ID. */\nexport async function getScenario(id: string): Promise<ScenarioSummary> {\n  return makeRequest(\n    \"GET\",\n    `/api/scenarios/${encodeURIComponent(id)}`\n  ) as Promise<ScenarioSummary>;\n}\n\n/** Creates a new scenario. */\nexport async function createScenario(data: {\n  name: string;\n  situation: string;\n  criteria?: string[];\n  labels?: string[];\n}): Promise<ScenarioSummary> {\n  return makeRequest(\"POST\", \"/api/scenarios\", data) as Promise<ScenarioSummary>;\n}\n\n/** Updates an existing scenario. */\nexport async function updateScenario(params: {\n  id: string;\n  name?: string;\n  situation?: string;\n  criteria?: string[];\n  labels?: string[];\n}): Promise<ScenarioSummary> {\n  const { id, ...data } = params;\n  return makeRequest(\n    \"PUT\",\n    `/api/scenarios/${encodeURIComponent(id)}`,\n    data\n  ) as Promise<ScenarioSummary>;\n}\n\n/** Archives (soft-deletes) a scenario. */\nexport async function archiveScenario(\n  id: string\n): Promise<ScenarioArchiveResponse> {\n  return makeRequest(\n    \"DELETE\",\n    `/api/scenarios/${encodeURIComponent(id)}`\n  ) as Promise<ScenarioArchiveResponse>;\n}\n"],"mappings":";;;;;AAoBA,eAAsB,gBAA4C;AAChE,SAAO,YAAY,OAAO,gBAAgB;AAC5C;AAGA,eAAsB,YAAY,IAAsC;AACtE,SAAO;AAAA,IACL;AAAA,IACA,kBAAkB,mBAAmB,EAAE,CAAC;AAAA,EAC1C;AACF;AAGA,eAAsB,eAAe,MAKR;AAC3B,SAAO,YAAY,QAAQ,kBAAkB,IAAI;AACnD;AAGA,eAAsB,eAAe,QAMR;AAC3B,QAAM,EAAE,IAAI,GAAG,KAAK,IAAI;AACxB,SAAO;AAAA,IACL;AAAA,IACA,kBAAkB,mBAAmB,EAAE,CAAC;AAAA,IACxC;AAAA,EACF;AACF;AAGA,eAAsB,gBACpB,IACkC;AAClC,SAAO;AAAA,IACL;AAAA,IACA,kBAAkB,mBAAmB,EAAE,CAAC;AAAA,EAC1C;AACF;","names":[]}

package/dist/{create-prompt-UBC537BJ.js → create-prompt-P35POKBW.js}

@@ -1,6 +1,6 @@
 import {
   createPrompt
-} from "./chunk-HOPTUDCZ.js";
+} from "./chunk-JVWDWL3J.js";
 import "./chunk-AAQNA53E.js";
 
 // src/tools/create-prompt.ts
@@ -19,4 +19,4 @@ async function handleCreatePrompt(params) {
 export {
   handleCreatePrompt
 };
-//# sourceMappingURL=create-prompt-UBC537BJ.js.map
+//# sourceMappingURL=create-prompt-P35POKBW.js.map

package/dist/create-scenario-3YRZVDYF.js

@@ -0,0 +1,26 @@
+import {
+  createScenario
+} from "./chunk-K2YFPOSD.js";
+import "./chunk-JVWDWL3J.js";
+import "./chunk-AAQNA53E.js";
+
+// src/tools/create-scenario.ts
+async function handleCreateScenario(params) {
+  const result = await createScenario(params);
+  const lines = [];
+  lines.push("Scenario created successfully!\n");
+  lines.push(`**ID**: ${result.id}`);
+  lines.push(`**Name**: ${result.name}`);
+  lines.push(`**Situation**: ${result.situation}`);
+  if (Array.isArray(result.criteria) && result.criteria.length > 0) {
+    lines.push(`**Criteria**: ${result.criteria.length} criteria`);
+  }
+  if (Array.isArray(result.labels) && result.labels.length > 0) {
+    lines.push(`**Labels**: ${result.labels.join(", ")}`);
+  }
+  return lines.join("\n");
+}
+export {
+  handleCreateScenario
+};
+//# sourceMappingURL=create-scenario-3YRZVDYF.js.map

package/dist/create-scenario-3YRZVDYF.js.map

@@ -0,0 +1 @@
+{"version":3,"sources":["../src/tools/create-scenario.ts"],"sourcesContent":["import { createScenario as apiCreateScenario } from \"../langwatch-api-scenarios.js\";\n\n/**\n * Handles the create_scenario MCP tool invocation.\n *\n * Creates a new scenario in the LangWatch project and returns a\n * confirmation with the created scenario's details.\n */\nexport async function handleCreateScenario(params: {\n  name: string;\n  situation: string;\n  criteria?: string[];\n  labels?: string[];\n}): Promise<string> {\n  const result = await apiCreateScenario(params);\n\n  const lines: string[] = [];\n  lines.push(\"Scenario created successfully!\\n\");\n  lines.push(`**ID**: ${result.id}`);\n  lines.push(`**Name**: ${result.name}`);\n  lines.push(`**Situation**: ${result.situation}`);\n  if (Array.isArray(result.criteria) && result.criteria.length > 0) {\n    lines.push(`**Criteria**: ${result.criteria.length} criteria`);\n  }\n  if (Array.isArray(result.labels) && result.labels.length > 0) {\n    lines.push(`**Labels**: ${result.labels.join(\", \")}`);\n  }\n\n  return lines.join(\"\\n\");\n}\n"],"mappings":";;;;;;;AAQA,eAAsB,qBAAqB,QAKvB;AAClB,QAAM,SAAS,MAAM,eAAkB,MAAM;AAE7C,QAAM,QAAkB,CAAC;AACzB,QAAM,KAAK,kCAAkC;AAC7C,QAAM,KAAK,WAAW,OAAO,EAAE,EAAE;AACjC,QAAM,KAAK,aAAa,OAAO,IAAI,EAAE;AACrC,QAAM,KAAK,kBAAkB,OAAO,SAAS,EAAE;AAC/C,MAAI,MAAM,QAAQ,OAAO,QAAQ,KAAK,OAAO,SAAS,SAAS,GAAG;AAChE,UAAM,KAAK,iBAAiB,OAAO,SAAS,MAAM,WAAW;AAAA,EAC/D;AACA,MAAI,MAAM,QAAQ,OAAO,MAAM,KAAK,OAAO,OAAO,SAAS,GAAG;AAC5D,UAAM,KAAK,eAAe,OAAO,OAAO,KAAK,IAAI,CAAC,EAAE;AAAA,EACtD;AAEA,SAAO,MAAM,KAAK,IAAI;AACxB;","names":[]}

package/dist/discover-scenario-schema-MEEEVND7.js

@@ -0,0 +1,65 @@
+// src/tools/discover-scenario-schema.ts
+function formatScenarioSchema() {
+  const lines = [];
+  lines.push("# Scenario Schema\n");
+  lines.push("## Fields\n");
+  lines.push(
+    '- **name** (required): A short, descriptive name (e.g., "billing dispute resolution", "password reset with 2FA unavailable")'
+  );
+  lines.push(
+    "- **situation** (required): The context that guides the user simulator \u2014 who the user is, what they want, and any constraints (see Writing a Good Situation below)"
+  );
+  lines.push(
+    "- **criteria** (array of strings): Pass/fail conditions a judge evaluates the agent against (see Writing Good Criteria below)"
+  );
+  lines.push(
+    '- **labels** (array of strings): Tags for organizing scenarios (e.g., "auth", "happy-path", "edge-case")'
+  );
+  lines.push("\n## Writing a Good Situation\n");
+  lines.push(
+    "The situation drives the user simulator. Include these elements:"
+  );
+  lines.push("- **Persona**: Who is the user? (e.g., a stressed small business owner, a confused teenager)");
+  lines.push("- **Emotional state**: How are they feeling? (e.g., frustrated, anxious, impatient)");
+  lines.push("- **Background/Context**: What happened before this conversation?");
+  lines.push("- **Intent**: What do they want to accomplish?");
+  lines.push("- **Constraints**: What limitations do they have? (e.g., no phone for 2FA, unfamiliar with technical terms)");
+  lines.push("\nExample:");
+  lines.push("```");
+  lines.push("User is a small business owner stressed about tax deadline.");
+  lines.push("They need help categorizing expenses but aren't familiar with");
+  lines.push("accounting terms. They appreciate patient explanations and examples.");
+  lines.push("They have a spreadsheet of transactions but aren't sure which");
+  lines.push("categories apply to their consulting business.");
+  lines.push("```");
+  lines.push("\n## Writing Good Criteria\n");
+  lines.push("Criteria are what the judge uses to pass or fail the agent. Each criterion should be:");
+  lines.push('- **Specific and testable** \u2014 not vague like "responds helpfully"');
+  lines.push("- **Behavioral** \u2014 describes what the agent should *do*, not how it works internally");
+  lines.push("- **Independent** \u2014 each criterion checks one thing");
+  lines.push("\nGood criteria patterns:");
+  lines.push(`- **Information gathering**: "Agent asks for the user's account number before proceeding"`);
+  lines.push('- **Safety/guardrails**: "Agent does not reveal internal system details or error stack traces"');
+  lines.push('- **Clarification**: "Agent asks clarifying questions before taking irreversible action"');
+  lines.push('- **Tone**: "Agent maintains a professional and empathetic tone throughout"');
+  lines.push('- **Completeness**: "Agent confirms the user understands the solution before ending"');
+  lines.push('- **Domain-specific**: "Agent recommends releasing a wild frog rather than keeping it as a pet"');
+  lines.push("\nAvoid vague criteria like:");
+  lines.push('- "Responds correctly" \u2014 correct how?');
+  lines.push('- "Is helpful" \u2014 helpful in what way?');
+  lines.push('- "Works well" \u2014 not testable');
+  lines.push("\n## Target Types\n");
+  lines.push("Scenarios can target different execution backends:");
+  lines.push("- **prompt**: Test a prompt template with variable substitution");
+  lines.push("- **http**: Test an HTTP endpoint (e.g., a deployed agent API)");
+  lines.push("- **code**: Test a code function directly");
+  lines.push("\n## Tips\n");
+  lines.push("- Start simple, then layer complexity (add constraints, edge cases)");
+  lines.push("- Test edge cases: user changes their mind, gives ambiguous input, makes mistakes");
+  lines.push("- Use `fetch_scenario_docs` for the full authoring guide and advanced patterns");
+  return lines.join("\n");
+}
+export {
+  formatScenarioSchema
+};
+//# sourceMappingURL=discover-scenario-schema-MEEEVND7.js.map

package/dist/discover-scenario-schema-MEEEVND7.js.map

@@ -0,0 +1 @@
+{"version":3,"sources":["../src/tools/discover-scenario-schema.ts"],"sourcesContent":["/**\n * Returns a human-readable description of the scenario schema,\n * including field descriptions, authoring guidance, and examples.\n */\nexport function formatScenarioSchema(): string {\n  const lines: string[] = [];\n\n  lines.push(\"# Scenario Schema\\n\");\n\n  lines.push(\"## Fields\\n\");\n  lines.push(\n    '- **name** (required): A short, descriptive name (e.g., \"billing dispute resolution\", \"password reset with 2FA unavailable\")',\n  );\n  lines.push(\n    \"- **situation** (required): The context that guides the user simulator — who the user is, what they want, and any constraints (see Writing a Good Situation below)\",\n  );\n  lines.push(\n    \"- **criteria** (array of strings): Pass/fail conditions a judge evaluates the agent against (see Writing Good Criteria below)\",\n  );\n  lines.push(\n    '- **labels** (array of strings): Tags for organizing scenarios (e.g., \"auth\", \"happy-path\", \"edge-case\")',\n  );\n\n  lines.push(\"\\n## Writing a Good Situation\\n\");\n  lines.push(\n    \"The situation drives the user simulator. Include these elements:\",\n  );\n  lines.push(\"- **Persona**: Who is the user? (e.g., a stressed small business owner, a confused teenager)\");\n  lines.push(\"- **Emotional state**: How are they feeling? (e.g., frustrated, anxious, impatient)\");\n  lines.push(\"- **Background/Context**: What happened before this conversation?\");\n  lines.push(\"- **Intent**: What do they want to accomplish?\");\n  lines.push(\"- **Constraints**: What limitations do they have? (e.g., no phone for 2FA, unfamiliar with technical terms)\");\n  lines.push(\"\\nExample:\");\n  lines.push(\"```\");\n  lines.push(\"User is a small business owner stressed about tax deadline.\");\n  lines.push(\"They need help categorizing expenses but aren't familiar with\");\n  lines.push(\"accounting terms. They appreciate patient explanations and examples.\");\n  lines.push(\"They have a spreadsheet of transactions but aren't sure which\");\n  lines.push(\"categories apply to their consulting business.\");\n  lines.push(\"```\");\n\n  lines.push(\"\\n## Writing Good Criteria\\n\");\n  lines.push(\"Criteria are what the judge uses to pass or fail the agent. Each criterion should be:\");\n  lines.push(\"- **Specific and testable** — not vague like \\\"responds helpfully\\\"\");\n  lines.push(\"- **Behavioral** — describes what the agent should *do*, not how it works internally\");\n  lines.push(\"- **Independent** — each criterion checks one thing\");\n  lines.push(\"\\nGood criteria patterns:\");\n  lines.push(\"- **Information gathering**: \\\"Agent asks for the user's account number before proceeding\\\"\");\n  lines.push(\"- **Safety/guardrails**: \\\"Agent does not reveal internal system details or error stack traces\\\"\");\n  lines.push(\"- **Clarification**: \\\"Agent asks clarifying questions before taking irreversible action\\\"\");\n  lines.push(\"- **Tone**: \\\"Agent maintains a professional and empathetic tone throughout\\\"\");\n  lines.push(\"- **Completeness**: \\\"Agent confirms the user understands the solution before ending\\\"\");\n  lines.push(\"- **Domain-specific**: \\\"Agent recommends releasing a wild frog rather than keeping it as a pet\\\"\");\n  lines.push(\"\\nAvoid vague criteria like:\");\n  lines.push('- \"Responds correctly\" — correct how?');\n  lines.push('- \"Is helpful\" — helpful in what way?');\n  lines.push('- \"Works well\" — not testable');\n\n  lines.push(\"\\n## Target Types\\n\");\n  lines.push(\"Scenarios can target different execution backends:\");\n  lines.push(\"- **prompt**: Test a prompt template with variable substitution\");\n  lines.push(\"- **http**: Test an HTTP endpoint (e.g., a deployed agent API)\");\n  lines.push(\"- **code**: Test a code function directly\");\n\n  lines.push(\"\\n## Tips\\n\");\n  lines.push(\"- Start simple, then layer complexity (add constraints, edge cases)\");\n  lines.push(\"- Test edge cases: user changes their mind, gives ambiguous input, makes mistakes\");\n  lines.push(\"- Use `fetch_scenario_docs` for the full authoring guide and advanced patterns\");\n\n  return lines.join(\"\\n\");\n}\n"],"mappings":";AAIO,SAAS,uBAA+B;AAC7C,QAAM,QAAkB,CAAC;AAEzB,QAAM,KAAK,qBAAqB;AAEhC,QAAM,KAAK,aAAa;AACxB,QAAM;AAAA,IACJ;AAAA,EACF;AACA,QAAM;AAAA,IACJ;AAAA,EACF;AACA,QAAM;AAAA,IACJ;AAAA,EACF;AACA,QAAM;AAAA,IACJ;AAAA,EACF;AAEA,QAAM,KAAK,iCAAiC;AAC5C,QAAM;AAAA,IACJ;AAAA,EACF;AACA,QAAM,KAAK,8FAA8F;AACzG,QAAM,KAAK,qFAAqF;AAChG,QAAM,KAAK,mEAAmE;AAC9E,QAAM,KAAK,gDAAgD;AAC3D,QAAM,KAAK,6GAA6G;AACxH,QAAM,KAAK,YAAY;AACvB,QAAM,KAAK,KAAK;AAChB,QAAM,KAAK,6DAA6D;AACxE,QAAM,KAAK,+DAA+D;AAC1E,QAAM,KAAK,sEAAsE;AACjF,QAAM,KAAK,+DAA+D;AAC1E,QAAM,KAAK,gDAAgD;AAC3D,QAAM,KAAK,KAAK;AAEhB,QAAM,KAAK,8BAA8B;AACzC,QAAM,KAAK,uFAAuF;AAClG,QAAM,KAAK,wEAAqE;AAChF,QAAM,KAAK,2FAAsF;AACjG,QAAM,KAAK,0DAAqD;AAChE,QAAM,KAAK,2BAA2B;AACtC,QAAM,KAAK,2FAA6F;AACxG,QAAM,KAAK,gGAAkG;AAC7G,QAAM,KAAK,0FAA4F;AACvG,QAAM,KAAK,6EAA+E;AAC1F,QAAM,KAAK,sFAAwF;AACnG,QAAM,KAAK,iGAAmG;AAC9G,QAAM,KAAK,8BAA8B;AACzC,QAAM,KAAK,4CAAuC;AAClD,QAAM,KAAK,4CAAuC;AAClD,QAAM,KAAK,oCAA+B;AAE1C,QAAM,KAAK,qBAAqB;AAChC,QAAM,KAAK,oDAAoD;AAC/D,QAAM,KAAK,iEAAiE;AAC5E,QAAM,KAAK,gEAAgE;AAC3E,QAAM,KAAK,2CAA2C;AAEtD,QAAM,KAAK,aAAa;AACxB,QAAM,KAAK,qEAAqE;AAChF,QAAM,KAAK,mFAAmF;AAC9F,QAAM,KAAK,gFAAgF;AAE3F,SAAO,MAAM,KAAK,IAAI;AACxB;","names":[]}

package/dist/{get-analytics-3IFTN6MY.js → get-analytics-BAVXTAPB.js}

@@ -3,7 +3,7 @@ import {
 } from "./chunk-ZXKLPC2E.js";
 import {
   getAnalyticsTimeseries
-} from "./chunk-HOPTUDCZ.js";
+} from "./chunk-JVWDWL3J.js";
 import "./chunk-AAQNA53E.js";
 
 // src/tools/get-analytics.ts
@@ -52,4 +52,4 @@ async function handleGetAnalytics(params) {
 export {
   handleGetAnalytics
 };
-//# sourceMappingURL=get-analytics-3IFTN6MY.js.map
+//# sourceMappingURL=get-analytics-BAVXTAPB.js.map

package/dist/{get-prompt-2ZB5B3QC.js → get-prompt-LKCPT26O.js}

@@ -1,6 +1,6 @@
 import {
   getPrompt
-} from "./chunk-HOPTUDCZ.js";
+} from "./chunk-JVWDWL3J.js";
 import "./chunk-AAQNA53E.js";
 
 // src/tools/get-prompt.ts
@@ -45,4 +45,4 @@ async function handleGetPrompt(params) {
 export {
   handleGetPrompt
 };
-//# sourceMappingURL=get-prompt-2ZB5B3QC.js.map
+//# sourceMappingURL=get-prompt-LKCPT26O.js.map

package/dist/get-scenario-3SCDW4Z6.js

@@ -0,0 +1,33 @@
+import {
+  getScenario
+} from "./chunk-K2YFPOSD.js";
+import "./chunk-JVWDWL3J.js";
+import "./chunk-AAQNA53E.js";
+
+// src/tools/get-scenario.ts
+async function handleGetScenario(params) {
+  const scenario = await getScenario(params.scenarioId);
+  if (params.format === "json") {
+    return JSON.stringify(scenario, null, 2);
+  }
+  const lines = [];
+  lines.push(`# Scenario: ${scenario.name}
+`);
+  lines.push(`**ID**: ${scenario.id}`);
+  lines.push(`**Situation**: ${scenario.situation}`);
+  if (Array.isArray(scenario.criteria) && scenario.criteria.length > 0) {
+    lines.push("\n## Criteria");
+    for (const criterion of scenario.criteria) {
+      lines.push(`- ${criterion}`);
+    }
+  }
+  if (Array.isArray(scenario.labels) && scenario.labels.length > 0) {
+    lines.push(`
+**Labels**: ${scenario.labels.join(", ")}`);
+  }
+  return lines.join("\n");
+}
+export {
+  handleGetScenario
+};
+//# sourceMappingURL=get-scenario-3SCDW4Z6.js.map

package/dist/get-scenario-3SCDW4Z6.js.map

@@ -0,0 +1 @@
+{"version":3,"sources":["../src/tools/get-scenario.ts"],"sourcesContent":["import { getScenario as apiGetScenario } from \"../langwatch-api-scenarios.js\";\n\n/**\n * Handles the get_scenario MCP tool invocation.\n *\n * Retrieves a specific scenario by ID and formats it as\n * AI-readable markdown or raw JSON.\n */\nexport async function handleGetScenario(params: {\n  scenarioId: string;\n  format?: \"digest\" | \"json\";\n}): Promise<string> {\n  const scenario = await apiGetScenario(params.scenarioId);\n\n  if (params.format === \"json\") {\n    return JSON.stringify(scenario, null, 2);\n  }\n\n  const lines: string[] = [];\n  lines.push(`# Scenario: ${scenario.name}\\n`);\n  lines.push(`**ID**: ${scenario.id}`);\n  lines.push(`**Situation**: ${scenario.situation}`);\n\n  if (Array.isArray(scenario.criteria) && scenario.criteria.length > 0) {\n    lines.push(\"\\n## Criteria\");\n    for (const criterion of scenario.criteria) {\n      lines.push(`- ${criterion}`);\n    }\n  }\n\n  if (Array.isArray(scenario.labels) && scenario.labels.length > 0) {\n    lines.push(`\\n**Labels**: ${scenario.labels.join(\", \")}`);\n  }\n\n  return lines.join(\"\\n\");\n}\n"],"mappings":";;;;;;;AAQA,eAAsB,kBAAkB,QAGpB;AAClB,QAAM,WAAW,MAAM,YAAe,OAAO,UAAU;AAEvD,MAAI,OAAO,WAAW,QAAQ;AAC5B,WAAO,KAAK,UAAU,UAAU,MAAM,CAAC;AAAA,EACzC;AAEA,QAAM,QAAkB,CAAC;AACzB,QAAM,KAAK,eAAe,SAAS,IAAI;AAAA,CAAI;AAC3C,QAAM,KAAK,WAAW,SAAS,EAAE,EAAE;AACnC,QAAM,KAAK,kBAAkB,SAAS,SAAS,EAAE;AAEjD,MAAI,MAAM,QAAQ,SAAS,QAAQ,KAAK,SAAS,SAAS,SAAS,GAAG;AACpE,UAAM,KAAK,eAAe;AAC1B,eAAW,aAAa,SAAS,UAAU;AACzC,YAAM,KAAK,KAAK,SAAS,EAAE;AAAA,IAC7B;AAAA,EACF;AAEA,MAAI,MAAM,QAAQ,SAAS,MAAM,KAAK,SAAS,OAAO,SAAS,GAAG;AAChE,UAAM,KAAK;AAAA,cAAiB,SAAS,OAAO,KAAK,IAAI,CAAC,EAAE;AAAA,EAC1D;AAEA,SAAO,MAAM,KAAK,IAAI;AACxB;","names":[]}

package/dist/{get-trace-7IXKKCJJ.js → get-trace-QFDWJ5D4.js}

@@ -1,6 +1,6 @@
 import {
   getTraceById
-} from "./chunk-HOPTUDCZ.js";
+} from "./chunk-JVWDWL3J.js";
 import "./chunk-AAQNA53E.js";
 
 // src/tools/get-trace.ts
@@ -47,4 +47,4 @@ ${result.formatted_trace}`);
 export {
   handleGetTrace
 };
-//# sourceMappingURL=get-trace-7IXKKCJJ.js.map
+//# sourceMappingURL=get-trace-QFDWJ5D4.js.map