@langwatch/mcp-server 0.8.0 → 0.10.0

package/dist/{chunk-B57MZ4WI.js → chunk-FVDAZ4XG.js}

@@ -1,3 +1,7 @@
+import {
+  LangWatchApiError,
+  makeRequest
+} from "./chunk-42ZHHHIJ.js";
 import {
   requireApiKey
 } from "./chunk-KOO5733S.js";
@@ -9,7 +13,7 @@ import { z as z2 } from "zod";
 // package.json
 var package_default = {
   name: "@langwatch/mcp-server",
-  version: "0.8.0",
+  version: "0.10.0",
   description: "An MCP server for Langwatch.",
   type: "module",
   main: "./dist/index.js",
@@ -60,7 +64,7 @@ var package_default = {
     "@types/node": "^25.6.0",
     "@types/yargs": "^17.0.33",
     "@typescript/native-preview": "7.0.0-dev.20250911.1",
-    ai: "^6.0.0",
+    ai: "^6.0.161",
     dotenv: "^17.2.2",
     esbuild: "^0.25.0",
     eslint: "^10.2.1",
@@ -87,8 +91,7 @@ var package_default = {
   },
   pnpm: {
     overrides: {
-      protobufjs: ">=7.5.6 <8.0.0 || >=8.0.2",
-      "fast-uri@<3.1.2": "3.1.2"
+      liquidjs: ">=10.26.0"
     }
   }
 };
@@ -119,6 +122,482 @@ var createDatasetSchema = z.object({
   columnTypes: z.array(datasetColumnDefinitionSchema).optional().describe("Column definitions for the dataset")
 });
 
+// src/tools/experiment-run-status.ts
+var INTERRUPTED_THRESHOLD_MS = 5 * 60 * 1e3;
+var deriveRunStatus = (timestamps, now = Date.now()) => {
+  if (timestamps.stoppedAt != null) return "stopped";
+  if (timestamps.finishedAt != null) return "completed";
+  if (timestamps.updatedAt != null && now - timestamps.updatedAt > INTERRUPTED_THRESHOLD_MS) {
+    return "interrupted";
+  }
+  return "running";
+};
+var isTerminalStatus = (status) => status === "completed" || status === "stopped";
+
+// src/tools/get-experiment-results.ts
+var DEFAULT_ROW_CAP = 50;
+var rowKey = (index, targetId) => `${index}:${targetId ?? ""}`;
+var summarizeEntry = (entry) => {
+  const candidates = ["input", "question", "query", "prompt", "user"];
+  for (const key of candidates) {
+    const value = entry[key];
+    if (typeof value === "string" && value.length > 0) {
+      return value.length > 80 ? `${value.slice(0, 77)}...` : value;
+    }
+  }
+  return "";
+};
+var isFailedEvaluation = (e) => e.status === "error" || e.passed === false;
+var isFailedRow = ({
+  entry,
+  evaluations
+}) => Boolean(entry.error) || evaluations.some(isFailedEvaluation);
+async function handleExperimentResults(params) {
+  const filter = params.filter ?? "all";
+  const evaluatorFilter = params.evaluator?.trim();
+  const limit = typeof params.limit === "number" && params.limit > 0 ? Math.min(params.limit, DEFAULT_ROW_CAP) : DEFAULT_ROW_CAP;
+  const search = new URLSearchParams();
+  if (params.experimentSlug) search.set("experimentSlug", params.experimentSlug);
+  const qs = search.toString() ? `?${search.toString()}` : "";
+  let results;
+  try {
+    results = await makeRequest(
+      "GET",
+      `/api/experiments/runs/${encodeURIComponent(params.runId)}/results${qs}`
+    );
+  } catch (error) {
+    const message = error instanceof Error ? error.message : String(error);
+    const status = error instanceof LangWatchApiError ? error.status : error && typeof error === "object" && "status" in error && typeof error.status === "number" ? error.status : void 0;
+    if (status === 404 || status === void 0 && /404|not found/i.test(message)) {
+      return [
+        `# Evaluation Results: ${params.runId}`,
+        "",
+        "**Status**: not found",
+        "",
+        `Could not load results for run \`${params.runId}\`. The run id may be incorrect, or its run state may have expired (Redis keeps it for 24h).`,
+        "",
+        "> Pass `experimentSlug` to load results for runs older than 24h. Discover the slug with `platform_experiment_list`, then `platform_experiment_list_runs` for the run ids."
+      ].join("\n");
+    }
+    throw error;
+  }
+  if (!results) {
+    return [
+      `# Evaluation Results: ${params.runId}`,
+      "",
+      "**Status**: not found",
+      "",
+      `Could not load results for run \`${params.runId}\`.`,
+      "",
+      "> Pass `experimentSlug` if the run is older than 24h. Discover the slug with `platform_experiment_list`, then `platform_experiment_list_runs` to confirm the run id."
+    ].join("\n");
+  }
+  const runStatus2 = deriveRunStatus(results.timestamps);
+  const partial = !isTerminalStatus(runStatus2);
+  const evaluationsByRow = /* @__PURE__ */ new Map();
+  for (const evaluation of results.evaluations) {
+    if (evaluatorFilter && evaluation.evaluator !== evaluatorFilter) continue;
+    const key = rowKey(evaluation.index, evaluation.targetId);
+    const list = evaluationsByRow.get(key) ?? [];
+    list.push(evaluation);
+    evaluationsByRow.set(key, list);
+  }
+  const evaluatorNames = Array.from(
+    new Set(
+      (evaluatorFilter ? results.evaluations.filter((e) => e.evaluator === evaluatorFilter) : results.evaluations).map((e) => e.evaluator)
+    )
+  );
+  let rows = results.dataset.map((entry) => ({
+    entry,
+    evaluations: evaluationsByRow.get(rowKey(entry.index, entry.targetId)) ?? []
+  }));
+  if (filter === "failed") {
+    rows = rows.filter(
+      (r) => isFailedRow({ entry: r.entry, evaluations: r.evaluations })
+    );
+  }
+  const totalMatching = rows.length;
+  const rowsForSummary = rows;
+  const truncated = rows.length > limit;
+  rows = rows.slice(0, limit);
+  const evaluatorAverages = /* @__PURE__ */ new Map();
+  for (const r of rowsForSummary) {
+    for (const e of r.evaluations) {
+      const stats = evaluatorAverages.get(e.evaluator) ?? {
+        sum: 0,
+        count: 0,
+        passed: 0,
+        failed: 0,
+        errored: 0
+      };
+      if (typeof e.score === "number") {
+        stats.sum += e.score;
+        stats.count += 1;
+      }
+      if (e.status === "error") stats.errored += 1;
+      else if (e.passed === true) stats.passed += 1;
+      else if (e.passed === false) stats.failed += 1;
+      evaluatorAverages.set(e.evaluator, stats);
+    }
+  }
+  const lines = [];
+  lines.push(`# Evaluation Results: ${results.runId}`);
+  lines.push("");
+  lines.push(`**Experiment**: ${results.experimentId}`);
+  lines.push(`**Status**: ${runStatus2}`);
+  if (typeof results.total === "number" && results.total > 0) {
+    lines.push(
+      `**Progress**: ${results.progress ?? results.dataset.length}/${results.total} rows`
+    );
+  }
+  lines.push(`**Total rows**: ${results.dataset.length}`);
+  lines.push(`**Total evaluations**: ${results.evaluations.length}`);
+  if (filter === "failed") {
+    lines.push(`**Filter**: failed only`);
+  }
+  if (evaluatorFilter) {
+    lines.push(`**Evaluator filter**: ${evaluatorFilter}`);
+  }
+  if (partial) {
+    lines.push("");
+    lines.push(
+      runStatus2 === "interrupted" ? "> These are partial results. The run never sent a finished/stopped marker and has had no updates recently, so it likely was interrupted before completing. The rows below are everything recorded so far." : "> These are partial results. The run is still in progress, so more rows may appear on a later call."
+    );
+  }
+  lines.push("");
+  if (evaluatorAverages.size > 0) {
+    lines.push("## Evaluator Summary");
+    lines.push("");
+    lines.push("| Evaluator | Avg Score | Passed | Failed | Errored |");
+    lines.push("| --- | --- | --- | --- | --- |");
+    for (const [name, stats] of evaluatorAverages) {
+      const avg = stats.count > 0 ? (stats.sum / stats.count).toFixed(3) : "\u2014";
+      lines.push(
+        `| ${name} | ${avg} | ${stats.passed} | ${stats.failed} | ${stats.errored} |`
+      );
+    }
+    lines.push("");
+  }
+  if (rows.length === 0) {
+    if (filter === "failed") {
+      lines.push("_No rows matched the filter._");
+    } else if (runStatus2 === "running") {
+      lines.push(
+        "_No rows recorded yet. The run is still in progress; call again shortly._"
+      );
+    } else if (runStatus2 === "interrupted") {
+      lines.push("_No rows were recorded before the run was interrupted._");
+    } else {
+      lines.push("_No rows recorded for this run._");
+    }
+    return lines.join("\n");
+  }
+  lines.push(
+    `## Rows (${rows.length}${truncated ? ` of ${totalMatching}` : ""})`
+  );
+  lines.push("");
+  for (const { entry, evaluations } of rows) {
+    const summary = summarizeEntry(entry.entry);
+    lines.push(`### Row #${entry.index}${summary ? ` \u2014 ${summary}` : ""}`);
+    if (entry.error) {
+      lines.push(`- **Error**: ${entry.error}`);
+    }
+    if (entry.traceId) {
+      lines.push(`- **Trace ID**: \`${entry.traceId}\``);
+    }
+    if (evaluations.length === 0) {
+      lines.push(
+        evaluatorNames.length === 0 ? "- _No evaluations recorded_" : "- _No evaluations recorded for this row_"
+      );
+    }
+    for (const e of evaluations) {
+      const parts = [`**${e.evaluator}**`];
+      if (e.status === "error") parts.push("status=error");
+      else if (typeof e.score === "number") parts.push(`score=${e.score.toFixed(3)}`);
+      if (typeof e.passed === "boolean") {
+        parts.push(`passed=${e.passed ? "yes" : "no"}`);
+      }
+      if (e.label) parts.push(`label=${e.label}`);
+      lines.push(`- ${parts.join(" \xB7 ")}`);
+      if (e.details) {
+        lines.push(`  - details: ${e.details}`);
+      }
+    }
+    lines.push("");
+  }
+  if (truncated) {
+    lines.push(
+      `> Output truncated to ${limit} rows of ${totalMatching} matching to protect the agent's context window. Pass \`limit\` to expand or filter further.`
+    );
+  }
+  return lines.join("\n");
+}
+
+// src/tools/list-experiment-runs.ts
+var DEFAULT_LIMIT = 25;
+var MAX_LIMIT = 100;
+var formatEpoch = (epoch) => {
+  if (!epoch || !Number.isFinite(epoch)) return "\u2014";
+  return new Date(epoch).toISOString().replace("T", " ").slice(0, 19) + " UTC";
+};
+var passRate = (evaluations) => {
+  const entries = Object.values(evaluations ?? {});
+  if (entries.length === 0) return "\u2014";
+  const passed = entries.filter((e) => typeof e.averagePassed === "number");
+  if (passed.length === 0) {
+    const scored = entries.filter((e) => typeof e.averageScore === "number");
+    if (scored.length === 0) return "\u2014";
+    const avg2 = scored.reduce((sum, e) => sum + (e.averageScore ?? 0), 0) / scored.length;
+    return `${avg2.toFixed(2)} avg`;
+  }
+  const avg = passed.reduce((sum, e) => sum + (e.averagePassed ?? 0), 0) / passed.length;
+  return `${(avg * 100).toFixed(0)}% pass`;
+};
+var runStatus = (run) => {
+  if (run.timestamps.stoppedAt) return "stopped";
+  if (run.timestamps.finishedAt) return "completed";
+  return "running";
+};
+async function handleExperimentListRuns(params) {
+  const requested = typeof params.limit === "number" && params.limit > 0 ? params.limit : DEFAULT_LIMIT;
+  const effectiveLimit = Math.min(requested, MAX_LIMIT);
+  const search = new URLSearchParams();
+  search.set("experimentSlug", params.experimentSlug);
+  search.set("pageSize", String(effectiveLimit));
+  let result;
+  try {
+    result = await makeRequest(
+      "GET",
+      `/api/experiments/runs?${search.toString()}`
+    );
+  } catch (error) {
+    const message = error instanceof Error ? error.message : String(error);
+    if (/404|not found/i.test(message)) {
+      return [
+        `# Evaluation Runs: ${params.experimentSlug}`,
+        "",
+        "**Status**: experiment not found",
+        "",
+        `No experiment with slug \`${params.experimentSlug}\` exists in this project.`,
+        "",
+        "> Run `platform_experiment_list` to discover available experiment slugs."
+      ].join("\n");
+    }
+    throw error;
+  }
+  if (result.runs.length === 0) {
+    return [
+      `# Evaluation Runs: ${result.experimentSlug}`,
+      "",
+      "_This experiment has no runs yet._",
+      "",
+      `> Trigger a run with \`platform_run_experiment\` (slug: \`${result.experimentSlug}\`) to populate this list.`
+    ].join("\n");
+  }
+  const lines = [];
+  lines.push(`# Evaluation Runs: ${result.experimentSlug}`);
+  lines.push("");
+  lines.push(
+    `Showing ${result.runs.length} of ${result.pagination.totalHits} run${result.pagination.totalHits === 1 ? "" : "s"}.`
+  );
+  if (result.pagination.hasMore) {
+    lines.push(
+      `> More runs exist beyond this limit. Re-run with a higher \`limit\` (max ${MAX_LIMIT}) if needed.`
+    );
+  }
+  lines.push("");
+  lines.push("| Run ID | Status | Started | Finished | Result |");
+  lines.push("|--------|--------|---------|----------|--------|");
+  for (const run of result.runs) {
+    lines.push(
+      `| \`${run.runId}\` | ${runStatus(run)} | ${formatEpoch(run.timestamps.createdAt)} | ${formatEpoch(run.timestamps.finishedAt ?? null)} | ${passRate(run.summary?.evaluations ?? {})} |`
+    );
+  }
+  lines.push("");
+  lines.push(
+    "> Use `platform_experiment_status <runId>` for live progress, or `platform_experiment_results <runId>` for per-row results."
+  );
+  return lines.join("\n");
+}
+
+// src/tools/list-experiments.ts
+var DEFAULT_LIMIT2 = 25;
+var MAX_LIMIT2 = 100;
+var formatTimestamp = (iso) => {
+  if (!iso) return "\u2014";
+  const d = new Date(iso);
+  if (Number.isNaN(d.getTime())) return iso;
+  return d.toISOString().replace("T", " ").slice(0, 19) + " UTC";
+};
+var escapeMarkdownTableCell = (value) => value.replace(/[\\|]/g, (c) => `\\${c}`).replace(/\n/g, " ");
+var formatInlineCode = (value) => `\`${value.replace(/[\\`|]/g, (c) => `\\${c}`).replace(/\n/g, " ")}\``;
+async function handleExperimentList(params) {
+  const requested = typeof params.limit === "number" && params.limit > 0 ? params.limit : DEFAULT_LIMIT2;
+  const effectiveLimit = Math.min(requested, MAX_LIMIT2);
+  const search = new URLSearchParams();
+  search.set("pageSize", String(effectiveLimit));
+  const result = await makeRequest(
+    "GET",
+    `/api/experiments?${search.toString()}`
+  );
+  if (result.experiments.length === 0) {
+    return [
+      "# Experiments",
+      "",
+      "_No experiments found in this project._",
+      "",
+      "> Create an experiment in the LangWatch dashboard, then re-run this tool to discover its slug."
+    ].join("\n");
+  }
+  const lines = [];
+  lines.push("# Experiments");
+  lines.push("");
+  lines.push(
+    `Showing ${result.experiments.length} of ${result.pagination.totalHits} experiment${result.pagination.totalHits === 1 ? "" : "s"}.`
+  );
+  if (result.pagination.hasMore) {
+    lines.push(
+      `> More experiments exist beyond this limit. Re-run with a higher \`limit\` (max ${MAX_LIMIT2}) if needed.`
+    );
+  }
+  lines.push("");
+  lines.push("| Slug | Name | Runs | Last Run |");
+  lines.push("|------|------|------|----------|");
+  for (const exp of result.experiments) {
+    const name = escapeMarkdownTableCell(exp.name ?? exp.slug);
+    lines.push(
+      `| ${formatInlineCode(exp.slug)} | ${name} | ${exp.runsCount} | ${formatTimestamp(exp.lastRunAt)} |`
+    );
+  }
+  lines.push("");
+  lines.push(
+    "> Use `platform_experiment_list_runs` with one of these slugs to see its runs."
+  );
+  return lines.join("\n");
+}
+
+// src/tools/run-experiment.ts
+async function handleRunExperiment(params) {
+  const result = await makeRequest(
+    "POST",
+    `/api/experiments/${encodeURIComponent(params.slug)}/run`
+  );
+  const lines = [];
+  lines.push(`Evaluation started!
+`);
+  lines.push(`**Run ID**: ${result.runId}`);
+  lines.push(`**Status**: ${result.status}`);
+  lines.push(`**Total cells**: ${result.total}`);
+  if (result.runUrl) lines.push(`**View at**: ${result.runUrl}`);
+  lines.push("");
+  lines.push(
+    `> Use \`platform_experiment_status\` with run ID "${result.runId}" to check progress.`
+  );
+  return lines.join("\n");
+}
+async function statusFromResults(params) {
+  const search = new URLSearchParams();
+  if (params.experimentSlug) search.set("experimentSlug", params.experimentSlug);
+  const qs = search.toString() ? `?${search.toString()}` : "";
+  let results;
+  try {
+    results = await makeRequest(
+      "GET",
+      `/api/experiments/runs/${encodeURIComponent(params.runId)}/results${qs}`
+    );
+  } catch (error) {
+    const code = error instanceof LangWatchApiError ? error.status : void 0;
+    const message = error instanceof Error ? error.message : String(error);
+    if (code === 404 || code === void 0 && /404|not found/i.test(message)) {
+      return null;
+    }
+    throw error;
+  }
+  const status = deriveRunStatus(results.timestamps);
+  const total = results.total ?? results.dataset?.length ?? 0;
+  const progress = results.progress ?? results.dataset?.length ?? 0;
+  const lines = [];
+  lines.push(`# Evaluation Run ${params.runId}
+`);
+  lines.push(`**Status**: ${status}`);
+  lines.push(`**Progress**: ${progress}/${total} cells`);
+  if (results.timestamps.createdAt) {
+    lines.push(
+      `**Started**: ${new Date(results.timestamps.createdAt).toISOString()}`
+    );
+  }
+  if (results.timestamps.finishedAt) {
+    lines.push(
+      `**Finished**: ${new Date(results.timestamps.finishedAt).toISOString()}`
+    );
+  }
+  if (results.timestamps.stoppedAt) {
+    lines.push(
+      `**Stopped**: ${new Date(results.timestamps.stoppedAt).toISOString()}`
+    );
+  }
+  lines.push("");
+  lines.push(
+    "> Use `platform_experiment_results` to fetch the per-row scores (partial results are available even while running)."
+  );
+  return lines.join("\n");
+}
+async function handleExperimentStatus(params) {
+  let status;
+  try {
+    status = await makeRequest(
+      "GET",
+      `/api/experiments/runs/${encodeURIComponent(params.runId)}`
+    );
+  } catch (error) {
+    const code = error instanceof LangWatchApiError ? error.status : void 0;
+    const message = error instanceof Error ? error.message : String(error);
+    if (code === 404 || /404|not found/i.test(message)) {
+      const fallback = await statusFromResults(params);
+      if (fallback) return fallback;
+      return [
+        `# Evaluation Run ${params.runId}`,
+        "",
+        "**Status**: not found",
+        "",
+        `Could not find run \`${params.runId}\`. SDK-logged runs and runs older than 24h are not in the live run-state and must be resolved by experiment slug.`,
+        "",
+        "> Pass `experimentSlug`: discover it with `platform_experiment_list`, then use `platform_experiment_list_runs` for the run ids. Or fetch the rows directly with `platform_experiment_results`."
+      ].join("\n");
+    }
+    throw error;
+  }
+  const lines = [];
+  lines.push(`# Evaluation Run ${status.runId}
+`);
+  lines.push(`**Status**: ${status.status}`);
+  lines.push(`**Progress**: ${status.progress}/${status.total} cells`);
+  if (status.startedAt) {
+    lines.push(`**Started**: ${new Date(status.startedAt).toISOString()}`);
+  }
+  if (status.finishedAt) {
+    lines.push(`**Finished**: ${new Date(status.finishedAt).toISOString()}`);
+  }
+  if (status.summary) {
+    lines.push("\n## Summary\n");
+    if (status.summary.completedCells !== void 0) {
+      lines.push(`**Completed**: ${status.summary.completedCells}`);
+    }
+    if (status.summary.failedCells) {
+      lines.push(`**Failed**: ${status.summary.failedCells}`);
+    }
+    if (status.summary.duration) {
+      lines.push(
+        `**Duration**: ${(status.summary.duration / 1e3).toFixed(1)}s`
+      );
+    }
+    if (status.summary.runUrl) {
+      lines.push(`**View results**: ${status.summary.runUrl}`);
+    }
+  }
+  return lines.join("\n");
+}
+
 // src/create-mcp-server.ts
 var modelSchema = z2.string().describe(
   'Model in "provider/model-name" format, e.g., "openai/gpt-4o", "anthropic/claude-sonnet-4-5-20250929"'
@@ -259,7 +738,7 @@ function registerTools(server) {
     },
     withToolLogging("search_traces", async (params) => {
       requireApiKey();
-      const { handleSearchTraces } = await import("./search-traces-ZCHNOMTX.js");
+      const { handleSearchTraces } = await import("./search-traces-3BBVYPC3.js");
       return {
         content: [{ type: "text", text: await handleSearchTraces(params) }]
       };
@@ -276,7 +755,7 @@ function registerTools(server) {
     },
     withToolLogging("get_trace", async (params) => {
       requireApiKey();
-      const { handleGetTrace } = await import("./get-trace-MUWEJM2Z.js");
+      const { handleGetTrace } = await import("./get-trace-MRBDS3HJ.js");
       return {
         content: [{ type: "text", text: await handleGetTrace(params) }]
       };
@@ -302,7 +781,7 @@ function registerTools(server) {
     },
     withToolLogging("get_analytics", async (params) => {
       requireApiKey();
-      const { handleGetAnalytics } = await import("./get-analytics-3W5CMU3F.js");
+      const { handleGetAnalytics } = await import("./get-analytics-GPVJX7SF.js");
       return {
         content: [{ type: "text", text: await handleGetAnalytics(params) }]
       };
@@ -336,7 +815,7 @@ NOTE: Prompts can be managed two ways. Determine which approach the user needs:
     },
     withToolLogging("platform_create_prompt", async (params) => {
       requireApiKey();
-      const { handleCreatePrompt } = await import("./create-prompt-NKRJ47YY.js");
+      const { handleCreatePrompt } = await import("./create-prompt-NGXKSP4X.js");
       return {
         content: [{ type: "text", text: await handleCreatePrompt(params) }]
       };
@@ -348,7 +827,7 @@ NOTE: Prompts can be managed two ways. Determine which approach the user needs:
     {},
     withToolLogging("platform_list_prompts", async () => {
       requireApiKey();
-      const { handleListPrompts } = await import("./list-prompts-7VTFWE2Z.js");
+      const { handleListPrompts } = await import("./list-prompts-VNXKHK7T.js");
       return {
         content: [{ type: "text", text: await handleListPrompts() }]
       };
@@ -372,7 +851,7 @@ NOTE: Prompts can be managed two ways. Determine which approach the user needs:
         };
       }
       requireApiKey();
-      const { handleGetPrompt } = await import("./get-prompt-XCSSNRKR.js");
+      const { handleGetPrompt } = await import("./get-prompt-5XAGWWBG.js");
       return {
         content: [{ type: "text", text: await handleGetPrompt(params) }]
       };
@@ -397,7 +876,7 @@ NOTE: Prompts can be managed two ways. Determine which approach the user needs:
     },
     withToolLogging("platform_update_prompt", async (params) => {
       requireApiKey();
-      const { handleUpdatePrompt } = await import("./update-prompt-3XEHECJ7.js");
+      const { handleUpdatePrompt } = await import("./update-prompt-ZU2DD6KM.js");
       return {
         content: [{ type: "text", text: await handleUpdatePrompt(params) }]
       };
@@ -413,7 +892,7 @@ NOTE: Prompts can be managed two ways. Determine which approach the user needs:
     },
     withToolLogging("platform_assign_prompt_tag", async (params) => {
       requireApiKey();
-      const { handleAssignPromptTag } = await import("./assign-prompt-tag-Y2B4CBBT.js");
+      const { handleAssignPromptTag } = await import("./assign-prompt-tag-FKCCZPML.js");
       return {
         content: [{ type: "text", text: await handleAssignPromptTag(params) }]
       };
@@ -425,7 +904,7 @@ NOTE: Prompts can be managed two ways. Determine which approach the user needs:
     {},
     withToolLogging("platform_list_prompt_tags", async () => {
       requireApiKey();
-      const { handleListPromptTags } = await import("./list-prompt-tags-PL5OISUA.js");
+      const { handleListPromptTags } = await import("./list-prompt-tags-Z2DD6FLQ.js");
       return {
         content: [{ type: "text", text: await handleListPromptTags() }]
       };
@@ -439,7 +918,7 @@ NOTE: Prompts can be managed two ways. Determine which approach the user needs:
     },
     withToolLogging("platform_create_prompt_tag", async (params) => {
       requireApiKey();
-      const { handleCreatePromptTag } = await import("./create-prompt-tag-S4JX3ULL.js");
+      const { handleCreatePromptTag } = await import("./create-prompt-tag-IUKYPVHH.js");
       return {
         content: [{ type: "text", text: await handleCreatePromptTag(params) }]
       };
@@ -454,7 +933,7 @@ NOTE: Prompts can be managed two ways. Determine which approach the user needs:
     },
     withToolLogging("platform_rename_prompt_tag", async (params) => {
       requireApiKey();
-      const { handleRenamePromptTag } = await import("./rename-prompt-tag-TURD6KYU.js");
+      const { handleRenamePromptTag } = await import("./rename-prompt-tag-A4RJUK7Z.js");
       return {
         content: [{ type: "text", text: await handleRenamePromptTag(params) }]
       };
@@ -468,7 +947,7 @@ NOTE: Prompts can be managed two ways. Determine which approach the user needs:
     },
     withToolLogging("platform_delete_prompt_tag", async (params) => {
       requireApiKey();
-      const { handleDeletePromptTag } = await import("./delete-prompt-tag-B5UKZL45.js");
+      const { handleDeletePromptTag } = await import("./delete-prompt-tag-I6HNMSVJ.js");
       return {
         content: [{ type: "text", text: await handleDeletePromptTag(params) }]
       };
@@ -496,7 +975,7 @@ NOTE: Scenarios can be created two ways. Determine which approach the user needs
     },
     withToolLogging("platform_create_scenario", async (params) => {
       requireApiKey();
-      const { handleCreateScenario } = await import("./create-scenario-BPRF75KF.js");
+      const { handleCreateScenario } = await import("./create-scenario-FNBXALH3.js");
       return {
         content: [
           { type: "text", text: await handleCreateScenario(params) }
@@ -514,7 +993,7 @@ NOTE: Scenarios can be created two ways. Determine which approach the user needs
     },
     withToolLogging("platform_list_scenarios", async (params) => {
       requireApiKey();
-      const { handleListScenarios } = await import("./list-scenarios-5BZWZ6ZN.js");
+      const { handleListScenarios } = await import("./list-scenarios-CFZHXL76.js");
       return {
         content: [
           { type: "text", text: await handleListScenarios(params) }
@@ -533,7 +1012,7 @@ NOTE: Scenarios can be created two ways. Determine which approach the user needs
     },
     withToolLogging("platform_get_scenario", async (params) => {
       requireApiKey();
-      const { handleGetScenario } = await import("./get-scenario-FNRKOY7E.js");
+      const { handleGetScenario } = await import("./get-scenario-7XBCDZWX.js");
       return {
         content: [{ type: "text", text: await handleGetScenario(params) }]
       };
@@ -551,7 +1030,7 @@ NOTE: Scenarios can be created two ways. Determine which approach the user needs
     },
     withToolLogging("platform_update_scenario", async (params) => {
       requireApiKey();
-      const { handleUpdateScenario } = await import("./update-scenario-DFPREYCD.js");
+      const { handleUpdateScenario } = await import("./update-scenario-3VUTPDRU.js");
       return {
         content: [
           { type: "text", text: await handleUpdateScenario(params) }
@@ -567,7 +1046,7 @@ NOTE: Scenarios can be created two ways. Determine which approach the user needs
     },
     withToolLogging("platform_archive_scenario", async (params) => {
       requireApiKey();
-      const { handleArchiveScenario } = await import("./archive-scenario-YKGLJ4MZ.js");
+      const { handleArchiveScenario } = await import("./archive-scenario-P5RRITRM.js");
       return {
         content: [
           { type: "text", text: await handleArchiveScenario(params) }
@@ -583,7 +1062,7 @@ NOTE: Scenarios can be created two ways. Determine which approach the user needs
     },
     withToolLogging("platform_list_suites", async (params) => {
       requireApiKey();
-      const { handleListSuites } = await import("./list-suites-RRQ2JX5Y.js");
+      const { handleListSuites } = await import("./list-suites-ZDOHAAXF.js");
       return {
         content: [{ type: "text", text: await handleListSuites(params) }]
       };
@@ -598,7 +1077,7 @@ NOTE: Scenarios can be created two ways. Determine which approach the user needs
     },
     withToolLogging("platform_get_suite", async (params) => {
       requireApiKey();
-      const { handleGetSuite } = await import("./get-suite-U5AOD5TC.js");
+      const { handleGetSuite } = await import("./get-suite-4DBVHFEM.js");
       return {
         content: [{ type: "text", text: await handleGetSuite(params) }]
       };
@@ -617,7 +1096,7 @@ NOTE: Scenarios can be created two ways. Determine which approach the user needs
     },
     withToolLogging("platform_create_suite", async (params) => {
       requireApiKey();
-      const { handleCreateSuite } = await import("./create-suite-4YN7D6K2.js");
+      const { handleCreateSuite } = await import("./create-suite-DJEPD4ZB.js");
       return {
         content: [{ type: "text", text: await handleCreateSuite(params) }]
       };
@@ -637,7 +1116,7 @@ NOTE: Scenarios can be created two ways. Determine which approach the user needs
     },
     withToolLogging("platform_update_suite", async (params) => {
       requireApiKey();
-      const { handleUpdateSuite } = await import("./update-suite-RATU5THO.js");
+      const { handleUpdateSuite } = await import("./update-suite-C6JFDQHL.js");
       return {
         content: [{ type: "text", text: await handleUpdateSuite(params) }]
       };
@@ -651,7 +1130,7 @@ NOTE: Scenarios can be created two ways. Determine which approach the user needs
     },
     withToolLogging("platform_run_suite", async (params) => {
       requireApiKey();
-      const { handleRunSuite } = await import("./run-suite-U7V432G2.js");
+      const { handleRunSuite } = await import("./run-suite-TH7LO5TJ.js");
       return {
         content: [{ type: "text", text: await handleRunSuite(params) }]
       };
@@ -665,7 +1144,7 @@ NOTE: Scenarios can be created two ways. Determine which approach the user needs
     },
     withToolLogging("platform_archive_suite", async (params) => {
       requireApiKey();
-      const { handleArchiveSuite } = await import("./archive-suite-QFNDO3Y3.js");
+      const { handleArchiveSuite } = await import("./archive-suite-NUUFD2EM.js");
       return {
         content: [{ type: "text", text: await handleArchiveSuite(params) }]
       };
@@ -682,7 +1161,7 @@ NOTE: Scenarios can be created two ways. Determine which approach the user needs
     },
     withToolLogging("platform_list_simulation_runs", async (params) => {
       requireApiKey();
-      const { handleListSimulationRuns } = await import("./list-simulation-runs-WCZ6RLPE.js");
+      const { handleListSimulationRuns } = await import("./list-simulation-runs-AM2BXLBW.js");
       return {
         content: [{ type: "text", text: await handleListSimulationRuns(params) }]
       };
@@ -697,7 +1176,7 @@ NOTE: Scenarios can be created two ways. Determine which approach the user needs
     },
     withToolLogging("platform_get_simulation_run", async (params) => {
       requireApiKey();
-      const { handleGetSimulationRun } = await import("./get-simulation-run-WOEKUF7K.js");
+      const { handleGetSimulationRun } = await import("./get-simulation-run-AV24FJWI.js");
       return {
         content: [{ type: "text", text: await handleGetSimulationRun(params) }]
       };
@@ -714,7 +1193,7 @@ NOTE: Scenarios can be created two ways. Determine which approach the user needs
     },
     withToolLogging("platform_create_evaluator", async (params) => {
       requireApiKey();
-      const { handleCreateEvaluator } = await import("./create-evaluator-2AU3PXFB.js");
+      const { handleCreateEvaluator } = await import("./create-evaluator-YFFQIR7E.js");
       return {
         content: [
           { type: "text", text: await handleCreateEvaluator(params) }
@@ -728,7 +1207,7 @@ NOTE: Scenarios can be created two ways. Determine which approach the user needs
     {},
     withToolLogging("platform_list_evaluators", async () => {
       requireApiKey();
-      const { handleListEvaluators } = await import("./list-evaluators-AKKRJ5MX.js");
+      const { handleListEvaluators } = await import("./list-evaluators-B6OSVAM2.js");
       return {
         content: [{ type: "text", text: await handleListEvaluators() }]
       };
@@ -742,7 +1221,7 @@ NOTE: Scenarios can be created two ways. Determine which approach the user needs
     },
     withToolLogging("platform_get_evaluator", async (params) => {
       requireApiKey();
-      const { handleGetEvaluator } = await import("./get-evaluator-QDUL5LXR.js");
+      const { handleGetEvaluator } = await import("./get-evaluator-PCM6XQAU.js");
       return {
         content: [{ type: "text", text: await handleGetEvaluator(params) }]
       };
@@ -760,7 +1239,7 @@ NOTE: Scenarios can be created two ways. Determine which approach the user needs
     },
     withToolLogging("platform_update_evaluator", async (params) => {
       requireApiKey();
-      const { handleUpdateEvaluator } = await import("./update-evaluator-LHVDXIZD.js");
+      const { handleUpdateEvaluator } = await import("./update-evaluator-6IJHFG2E.js");
       return {
         content: [
           { type: "text", text: await handleUpdateEvaluator(params) }
@@ -776,7 +1255,7 @@ NOTE: Scenarios can be created two ways. Determine which approach the user needs
     },
     withToolLogging("platform_delete_evaluator", async (params) => {
       requireApiKey();
-      const { handleDeleteEvaluator } = await import("./delete-evaluator-F7QLXDHU.js");
+      const { handleDeleteEvaluator } = await import("./delete-evaluator-7ZATLJOA.js");
       return {
         content: [{ type: "text", text: await handleDeleteEvaluator(params) }]
       };
@@ -797,7 +1276,7 @@ NOTE: Scenarios can be created two ways. Determine which approach the user needs
     },
     withToolLogging("platform_set_model_provider", async (params) => {
       requireApiKey();
-      const { handleSetModelProvider } = await import("./set-model-provider-6CFHMCNE.js");
+      const { handleSetModelProvider } = await import("./set-model-provider-WOC5FNMV.js");
       return {
         content: [
           { type: "text", text: await handleSetModelProvider(params) }
@@ -811,7 +1290,7 @@ NOTE: Scenarios can be created two ways. Determine which approach the user needs
     {},
     withToolLogging("platform_list_model_providers", async () => {
       requireApiKey();
-      const { handleListModelProviders } = await import("./list-model-providers-DVFEHFHZ.js");
+      const { handleListModelProviders } = await import("./list-model-providers-NL6IP4UZ.js");
       return {
         content: [{ type: "text", text: await handleListModelProviders() }]
       };
@@ -823,7 +1302,7 @@ NOTE: Scenarios can be created two ways. Determine which approach the user needs
     {},
     withToolLogging("platform_list_agents", async () => {
       requireApiKey();
-      const { handleListAgents } = await import("./list-agents-3HGIBJFM.js");
+      const { handleListAgents } = await import("./list-agents-NECO7MPI.js");
       return {
         content: [{ type: "text", text: await handleListAgents() }]
       };
@@ -837,7 +1316,7 @@ NOTE: Scenarios can be created two ways. Determine which approach the user needs
     },
     withToolLogging("platform_get_agent", async (params) => {
       requireApiKey();
-      const { handleGetAgent } = await import("./get-agent-YEVICUX4.js");
+      const { handleGetAgent } = await import("./get-agent-3KDWCP5N.js");
       return {
         content: [{ type: "text", text: await handleGetAgent(params) }]
       };
@@ -853,7 +1332,7 @@ NOTE: Scenarios can be created two ways. Determine which approach the user needs
     },
     withToolLogging("platform_create_agent", async (params) => {
       requireApiKey();
-      const { handleCreateAgent } = await import("./create-agent-YRSDH7P7.js");
+      const { handleCreateAgent } = await import("./create-agent-K5Q7BFGZ.js");
       const parsedConfig = params.config ? JSON.parse(params.config) : void 0;
       return {
         content: [{ type: "text", text: await handleCreateAgent({ ...params, config: parsedConfig }) }]
@@ -871,7 +1350,7 @@ NOTE: Scenarios can be created two ways. Determine which approach the user needs
     },
     withToolLogging("platform_update_agent", async (params) => {
       requireApiKey();
-      const { handleUpdateAgent } = await import("./update-agent-UE6UWLZP.js");
+      const { handleUpdateAgent } = await import("./update-agent-YMEV5MTL.js");
       const parsedConfig = params.config ? JSON.parse(params.config) : void 0;
       return {
         content: [{ type: "text", text: await handleUpdateAgent({ ...params, config: parsedConfig }) }]
@@ -886,7 +1365,7 @@ NOTE: Scenarios can be created two ways. Determine which approach the user needs
     },
     withToolLogging("platform_delete_agent", async (params) => {
       requireApiKey();
-      const { handleDeleteAgent } = await import("./delete-agent-Y72R2YQA.js");
+      const { handleDeleteAgent } = await import("./delete-agent-TJI5XBBM.js");
       return {
         content: [{ type: "text", text: await handleDeleteAgent(params) }]
       };
@@ -901,7 +1380,7 @@ NOTE: Scenarios can be created two ways. Determine which approach the user needs
     },
     withToolLogging("platform_run_agent", async (params) => {
       requireApiKey();
-      const { handleRunAgent } = await import("./run-agent-BEJ4K4CM.js");
+      const { handleRunAgent } = await import("./run-agent-JFNDY7MR.js");
       return {
         content: [{ type: "text", text: await handleRunAgent(params) }]
       };
@@ -913,7 +1392,7 @@ NOTE: Scenarios can be created two ways. Determine which approach the user needs
     {},
     withToolLogging("platform_list_dashboards", async () => {
       requireApiKey();
-      const { handleListDashboards } = await import("./list-dashboards-VQ2LH666.js");
+      const { handleListDashboards } = await import("./list-dashboards-DVPVVVKW.js");
       return {
         content: [{ type: "text", text: await handleListDashboards() }]
       };
@@ -927,7 +1406,7 @@ NOTE: Scenarios can be created two ways. Determine which approach the user needs
     },
     withToolLogging("platform_get_dashboard", async (params) => {
       requireApiKey();
-      const { handleGetDashboard } = await import("./get-dashboard-NCQHN6MW.js");
+      const { handleGetDashboard } = await import("./get-dashboard-YPTIE5YY.js");
       return {
         content: [{ type: "text", text: await handleGetDashboard(params) }]
       };
@@ -941,7 +1420,7 @@ NOTE: Scenarios can be created two ways. Determine which approach the user needs
     },
     withToolLogging("platform_create_dashboard", async (params) => {
       requireApiKey();
-      const { handleCreateDashboard } = await import("./create-dashboard-CORBR4YQ.js");
+      const { handleCreateDashboard } = await import("./create-dashboard-G4QA5AXP.js");
       return {
         content: [{ type: "text", text: await handleCreateDashboard(params) }]
       };
@@ -955,7 +1434,7 @@ NOTE: Scenarios can be created two ways. Determine which approach the user needs
     },
     withToolLogging("platform_delete_dashboard", async (params) => {
       requireApiKey();
-      const { handleDeleteDashboard } = await import("./delete-dashboard-LPWT36KN.js");
+      const { handleDeleteDashboard } = await import("./delete-dashboard-N5SZJIR3.js");
       return {
         content: [{ type: "text", text: await handleDeleteDashboard(params) }]
       };
@@ -970,7 +1449,7 @@ NOTE: Scenarios can be created two ways. Determine which approach the user needs
     },
     withToolLogging("platform_rename_dashboard", async (params) => {
       requireApiKey();
-      const { renameDashboard } = await import("./langwatch-api-dashboards-VBWWXV3Y.js");
+      const { renameDashboard } = await import("./langwatch-api-dashboards-WARLI6NF.js");
       const result = await renameDashboard(params.id, { name: params.name });
       return {
         content: [{ type: "text", text: `Dashboard "${result.name}" renamed successfully (ID: ${result.id}).` }]
@@ -983,7 +1462,7 @@ NOTE: Scenarios can be created two ways. Determine which approach the user needs
     {},
     withToolLogging("platform_list_workflows", async () => {
       requireApiKey();
-      const { handleListWorkflows } = await import("./list-workflows-UQB5FCX5.js");
+      const { handleListWorkflows } = await import("./list-workflows-WBRBKLIT.js");
       return {
         content: [{ type: "text", text: await handleListWorkflows() }]
       };
@@ -997,7 +1476,7 @@ NOTE: Scenarios can be created two ways. Determine which approach the user needs
     },
     withToolLogging("platform_get_workflow", async (params) => {
       requireApiKey();
-      const { handleGetWorkflow } = await import("./get-workflow-SOA7MT7L.js");
+      const { handleGetWorkflow } = await import("./get-workflow-UR4Y6UFX.js");
       return {
         content: [{ type: "text", text: await handleGetWorkflow(params) }]
       };
@@ -1011,7 +1490,7 @@ NOTE: Scenarios can be created two ways. Determine which approach the user needs
     },
     withToolLogging("platform_delete_workflow", async (params) => {
       requireApiKey();
-      const { handleDeleteWorkflow } = await import("./delete-workflow-4DGUDIDE.js");
+      const { handleDeleteWorkflow } = await import("./delete-workflow-QEO74JQR.js");
       return {
         content: [{ type: "text", text: await handleDeleteWorkflow(params) }]
       };
@@ -1026,7 +1505,7 @@ NOTE: Scenarios can be created two ways. Determine which approach the user needs
     },
     withToolLogging("platform_run_workflow", async (params) => {
       requireApiKey();
-      const { handleRunWorkflow } = await import("./run-workflow-5WX7DK5T.js");
+      const { handleRunWorkflow } = await import("./run-workflow-CK3JP5U2.js");
       return {
         content: [{ type: "text", text: await handleRunWorkflow(params) }]
       };
@@ -1040,7 +1519,7 @@ NOTE: Scenarios can be created two ways. Determine which approach the user needs
     },
     withToolLogging("platform_list_annotations", async (params) => {
       requireApiKey();
-      const { handleListAnnotations } = await import("./list-annotations-KWXYVEPI.js");
+      const { handleListAnnotations } = await import("./list-annotations-3OJEO774.js");
       return {
         content: [{ type: "text", text: await handleListAnnotations(params) }]
       };
@@ -1057,7 +1536,7 @@ NOTE: Scenarios can be created two ways. Determine which approach the user needs
     },
     withToolLogging("platform_create_annotation", async (params) => {
       requireApiKey();
-      const { handleCreateAnnotation } = await import("./create-annotation-FYQNKMVJ.js");
+      const { handleCreateAnnotation } = await import("./create-annotation-KN3GE3DK.js");
       return {
         content: [{ type: "text", text: await handleCreateAnnotation(params) }]
       };
@@ -1071,7 +1550,7 @@ NOTE: Scenarios can be created two ways. Determine which approach the user needs
     },
     withToolLogging("platform_delete_annotation", async (params) => {
       requireApiKey();
-      const { handleDeleteAnnotation } = await import("./delete-annotation-UHWDMHSH.js");
+      const { handleDeleteAnnotation } = await import("./delete-annotation-KRP53TUU.js");
       return {
         content: [{ type: "text", text: await handleDeleteAnnotation(params) }]
       };
@@ -1085,7 +1564,7 @@ NOTE: Scenarios can be created two ways. Determine which approach the user needs
     },
     withToolLogging("platform_get_annotation", async (params) => {
       requireApiKey();
-      const { getAnnotation } = await import("./langwatch-api-annotations-Q57NDUVR.js");
+      const { getAnnotation } = await import("./langwatch-api-annotations-ZJDRB4JH.js");
       const annotation = await getAnnotation(params.id);
       const lines = [
         `**ID**: ${annotation.id}`,
@@ -1107,7 +1586,7 @@ NOTE: Scenarios can be created two ways. Determine which approach the user needs
     },
     withToolLogging("platform_list_triggers", async (params) => {
       requireApiKey();
-      const { listTriggers } = await import("./langwatch-api-triggers-ME33QE26.js");
+      const { listTriggers } = await import("./langwatch-api-triggers-5SX6R2WU.js");
       const triggers = await listTriggers();
       if (params.format === "json") {
         return { content: [{ type: "text", text: JSON.stringify(triggers, null, 2) }] };
@@ -1140,7 +1619,7 @@ NOTE: Scenarios can be created two ways. Determine which approach the user needs
     },
     withToolLogging("platform_create_trigger", async (params) => {
       requireApiKey();
-      const { createTrigger } = await import("./langwatch-api-triggers-ME33QE26.js");
+      const { createTrigger } = await import("./langwatch-api-triggers-5SX6R2WU.js");
       let filters = {};
       if (params.filters) {
         try {
@@ -1171,7 +1650,7 @@ NOTE: Scenarios can be created two ways. Determine which approach the user needs
     },
     withToolLogging("platform_update_trigger", async (params) => {
       requireApiKey();
-      const { updateTrigger } = await import("./langwatch-api-triggers-ME33QE26.js");
+      const { updateTrigger } = await import("./langwatch-api-triggers-5SX6R2WU.js");
       const trigger = await updateTrigger(params);
       return { content: [{ type: "text", text: `Trigger "${trigger.name}" updated (active: ${trigger.active}).` }] };
     })
@@ -1184,7 +1663,7 @@ NOTE: Scenarios can be created two ways. Determine which approach the user needs
     },
     withToolLogging("platform_delete_trigger", async (params) => {
       requireApiKey();
-      const { deleteTrigger } = await import("./langwatch-api-triggers-ME33QE26.js");
+      const { deleteTrigger } = await import("./langwatch-api-triggers-5SX6R2WU.js");
       const result = await deleteTrigger(params.id);
       return { content: [{ type: "text", text: `Trigger ${result.id} deleted.` }] };
     })
@@ -1195,7 +1674,7 @@ NOTE: Scenarios can be created two ways. Determine which approach the user needs
     {},
     withToolLogging("platform_list_monitors", async () => {
       requireApiKey();
-      const { listMonitors } = await import("./langwatch-api-monitors-L3IY7FDT.js");
+      const { listMonitors } = await import("./langwatch-api-monitors-4XKXHYCG.js");
       const monitors = await listMonitors();
       if (monitors.length === 0) {
         return { content: [{ type: "text", text: "No monitors found." }] };
@@ -1222,7 +1701,7 @@ ${lines.join("\n")}`
     },
     withToolLogging("platform_get_monitor", async (params) => {
       requireApiKey();
-      const { getMonitor } = await import("./langwatch-api-monitors-L3IY7FDT.js");
+      const { getMonitor } = await import("./langwatch-api-monitors-4XKXHYCG.js");
       const m = await getMonitor(params.id);
       const lines = [
         `**${m.name}** (${m.id})`,
@@ -1249,7 +1728,7 @@ ${lines.join("\n")}`
     },
     withToolLogging("platform_create_monitor", async (params) => {
       requireApiKey();
-      const { createMonitor } = await import("./langwatch-api-monitors-L3IY7FDT.js");
+      const { createMonitor } = await import("./langwatch-api-monitors-4XKXHYCG.js");
       const monitor = await createMonitor(params);
       return {
         content: [
@@ -1273,7 +1752,7 @@ ${lines.join("\n")}`
     },
     withToolLogging("platform_update_monitor", async (params) => {
       requireApiKey();
-      const { updateMonitor } = await import("./langwatch-api-monitors-L3IY7FDT.js");
+      const { updateMonitor } = await import("./langwatch-api-monitors-4XKXHYCG.js");
       const monitor = await updateMonitor(params);
       return {
         content: [
@@ -1293,7 +1772,7 @@ ${lines.join("\n")}`
     },
     withToolLogging("platform_delete_monitor", async (params) => {
       requireApiKey();
-      const { deleteMonitor } = await import("./langwatch-api-monitors-L3IY7FDT.js");
+      const { deleteMonitor } = await import("./langwatch-api-monitors-4XKXHYCG.js");
       const result = await deleteMonitor(params.id);
       return {
         content: [
@@ -1308,7 +1787,7 @@ ${lines.join("\n")}`
     {},
     withToolLogging("platform_list_secrets", async () => {
       requireApiKey();
-      const { listSecrets } = await import("./langwatch-api-secrets-S5HHF3XR.js");
+      const { listSecrets } = await import("./langwatch-api-secrets-6SHK5G7Q.js");
       const secrets = await listSecrets();
       if (secrets.length === 0) {
         return { content: [{ type: "text", text: "No secrets found." }] };
@@ -1336,7 +1815,7 @@ ${lines.join("\n")}`
     },
     withToolLogging("platform_create_secret", async (params) => {
       requireApiKey();
-      const { createSecret } = await import("./langwatch-api-secrets-S5HHF3XR.js");
+      const { createSecret } = await import("./langwatch-api-secrets-6SHK5G7Q.js");
       const secret = await createSecret({ name: params.name, value: params.value });
       return {
         content: [
@@ -1357,7 +1836,7 @@ ${lines.join("\n")}`
     },
     withToolLogging("platform_update_secret", async (params) => {
       requireApiKey();
-      const { updateSecret } = await import("./langwatch-api-secrets-S5HHF3XR.js");
+      const { updateSecret } = await import("./langwatch-api-secrets-6SHK5G7Q.js");
       const secret = await updateSecret({ id: params.id, value: params.value });
       return {
         content: [
@@ -1377,36 +1856,95 @@ ${lines.join("\n")}`
     },
     withToolLogging("platform_delete_secret", async (params) => {
       requireApiKey();
-      const { deleteSecret } = await import("./langwatch-api-secrets-S5HHF3XR.js");
+      const { deleteSecret } = await import("./langwatch-api-secrets-6SHK5G7Q.js");
       const result = await deleteSecret(params.id);
       return { content: [{ type: "text", text: `Secret ${result.id} deleted.` }] };
     })
   );
   server.tool(
-    "platform_run_evaluation",
-    "Start an evaluation run by slug. Returns a run ID for polling status.",
+    "platform_run_experiment",
+    "Start an experiment run by slug. Returns a run ID for polling status.",
+    {
+      slug: z2.string().describe("The experiment slug to run")
+    },
+    withToolLogging("platform_run_experiment", async (params) => {
+      requireApiKey();
+      return {
+        content: [{ type: "text", text: await handleRunExperiment(params) }]
+      };
+    })
+  );
+  server.tool(
+    "platform_experiment_status",
+    "Check the status of an experiment run (running, completed, stopped, or interrupted) plus its progress. Works for both platform-run and SDK-logged experiments; pass experimentSlug for SDK runs or runs older than 24h.",
+    {
+      runId: z2.string().describe("The run ID returned from platform_run_experiment"),
+      experimentSlug: z2.string().optional().describe(
+        "Experiment slug \u2014 required for SDK-logged runs and runs older than 24h (the Redis run-state they would otherwise be read from is absent/expired). Use platform_experiment_list_runs to find the slug."
+      )
+    },
+    withToolLogging("platform_experiment_status", async (params) => {
+      requireApiKey();
+      return {
+        content: [{ type: "text", text: await handleExperimentStatus(params) }]
+      };
+    })
+  );
+  server.tool(
+    "platform_experiment_list",
+    "List experiments configured in the LangWatch project. Returns a markdown table with each experiment's slug, name, last-run timestamp, and run count. Use this to discover experiment slugs before drilling into runs with platform_experiment_list_runs. Output is capped (default 25, max 100) to protect the agent's context window.",
+    {
+      limit: z2.number().int().positive().max(100).optional().describe(
+        "Maximum number of experiments to include (default 25, hard-capped at 100)"
+      )
+    },
+    withToolLogging("platform_experiment_list", async (params) => {
+      requireApiKey();
+      return {
+        content: [{ type: "text", text: await handleExperimentList(params) }]
+      };
+    })
+  );
+  server.tool(
+    "platform_experiment_list_runs",
+    "List experiment runs for a specific experiment slug. Returns a markdown table with each run's id, status, started/finished timestamps, and pass-rate summary. Use this to discover run ids before calling platform_experiment_results. Output is capped (default 25, max 100) to protect the agent's context window.",
     {
-      slug: z2.string().describe("The evaluation slug to run")
+      experimentSlug: z2.string().describe(
+        "Experiment slug from platform_experiment_list (e.g. 'checkout-flow')"
+      ),
+      limit: z2.number().int().positive().max(100).optional().describe(
+        "Maximum number of runs to include (default 25, hard-capped at 100)"
+      )
     },
-    withToolLogging("platform_run_evaluation", async (params) => {
+    withToolLogging("platform_experiment_list_runs", async (params) => {
       requireApiKey();
-      const { handleRunEvaluation } = await import("./run-evaluation-AJ7WZNQD.js");
       return {
-        content: [{ type: "text", text: await handleRunEvaluation(params) }]
+        content: [{ type: "text", text: await handleExperimentListRuns(params) }]
       };
     })
   );
   server.tool(
-    "platform_evaluation_status",
-    "Check the status of an evaluation run. Returns progress and summary when completed.",
+    "platform_experiment_results",
+    "Fetch per-row results for an experiment run so you can debug evaluator scores and missed rows. Serves partial results for runs that are still running or were interrupted (it does not require a completed run), with the run status noted in the output. Returns a markdown report: per-evaluator averages plus row-by-row scores and failure details. Output is capped at 50 rows to protect the agent's context window \u2014 narrow with `filter: 'failed'` or `evaluator` to see what matters.",
     {
-      runId: z2.string().describe("The run ID returned from platform_run_evaluation")
+      runId: z2.string().describe("The run ID returned from platform_run_experiment"),
+      experimentSlug: z2.string().optional().describe(
+        "Experiment slug \u2014 required when fetching results for runs older than 24h (Redis state may have expired). Use platform_experiment_list_runs to find the slug."
+      ),
+      filter: z2.enum(["all", "failed"]).optional().describe(
+        "Filter rows: 'failed' shows only rows that errored or failed an evaluation; 'all' (default) shows everything"
+      ),
+      evaluator: z2.string().optional().describe("Show only this evaluator's results (matched on `evaluator` field)"),
+      limit: z2.number().int().positive().max(50).optional().describe(
+        "Maximum number of rows to include in the markdown output (default 50, hard-capped at 50)."
+      )
     },
-    withToolLogging("platform_evaluation_status", async (params) => {
+    withToolLogging("platform_experiment_results", async (params) => {
       requireApiKey();
-      const { handleEvaluationStatus } = await import("./run-evaluation-AJ7WZNQD.js");
       return {
-        content: [{ type: "text", text: await handleEvaluationStatus(params) }]
+        content: [
+          { type: "text", text: await handleExperimentResults(params) }
+        ]
       };
     })
   );
@@ -1421,7 +1959,7 @@ ${lines.join("\n")}`
     async (params) => {
       const { requireApiKey: requireApiKey2 } = await import("./config.js");
       requireApiKey2();
-      const { handleListDatasets } = await import("./list-datasets-I4A2LT2S.js");
+      const { handleListDatasets } = await import("./list-datasets-WUAE5HGX.js");
       return {
         content: [{ type: "text", text: await handleListDatasets(params) }]
       };
@@ -1439,7 +1977,7 @@ ${lines.join("\n")}`
     async (params) => {
       const { requireApiKey: requireApiKey2 } = await import("./config.js");
       requireApiKey2();
-      const { handleGetDataset } = await import("./get-dataset-KBK6PH36.js");
+      const { handleGetDataset } = await import("./get-dataset-6DQFYITM.js");
       return {
         content: [{ type: "text", text: await handleGetDataset(params) }]
       };
@@ -1452,7 +1990,7 @@ ${lines.join("\n")}`
     async (params) => {
       const { requireApiKey: requireApiKey2 } = await import("./config.js");
       requireApiKey2();
-      const { handleCreateDataset } = await import("./create-dataset-ZC63T6BB.js");
+      const { handleCreateDataset } = await import("./create-dataset-WQSC37N7.js");
       return {
         content: [{ type: "text", text: await handleCreateDataset(params) }]
       };
@@ -1469,7 +2007,7 @@ ${lines.join("\n")}`
     async (params) => {
       const { requireApiKey: requireApiKey2 } = await import("./config.js");
       requireApiKey2();
-      const { handleUpdateDataset } = await import("./update-dataset-4XQZZE4S.js");
+      const { handleUpdateDataset } = await import("./update-dataset-EYT5POFM.js");
       return {
         content: [{ type: "text", text: await handleUpdateDataset(params) }]
       };
@@ -1484,7 +2022,7 @@ ${lines.join("\n")}`
     async (params) => {
       const { requireApiKey: requireApiKey2 } = await import("./config.js");
       requireApiKey2();
-      const { handleDeleteDataset } = await import("./delete-dataset-GGPQRHKX.js");
+      const { handleDeleteDataset } = await import("./delete-dataset-KDUOHL2P.js");
       return {
         content: [{ type: "text", text: await handleDeleteDataset(params) }]
       };
@@ -1502,7 +2040,7 @@ ${lines.join("\n")}`
     async (params) => {
       const { requireApiKey: requireApiKey2 } = await import("./config.js");
       requireApiKey2();
-      const { handleListDatasetRecords } = await import("./list-dataset-records-RNRGROJP.js");
+      const { handleListDatasetRecords } = await import("./list-dataset-records-I4BYORF5.js");
       return {
         content: [
           { type: "text", text: await handleListDatasetRecords(params) }
@@ -1520,7 +2058,7 @@ ${lines.join("\n")}`
     async (params) => {
       const { requireApiKey: requireApiKey2 } = await import("./config.js");
       requireApiKey2();
-      const { handleCreateDatasetRecords } = await import("./create-dataset-records-ILVNVKVR.js");
+      const { handleCreateDatasetRecords } = await import("./create-dataset-records-B4RL3FLC.js");
       return {
         content: [
           { type: "text", text: await handleCreateDatasetRecords(params) }
@@ -1539,7 +2077,7 @@ ${lines.join("\n")}`
     async (params) => {
       const { requireApiKey: requireApiKey2 } = await import("./config.js");
       requireApiKey2();
-      const { handleUpdateDatasetRecord } = await import("./update-dataset-record-GNNM5TKM.js");
+      const { handleUpdateDatasetRecord } = await import("./update-dataset-record-VFF5JTX4.js");
       return {
         content: [
           { type: "text", text: await handleUpdateDatasetRecord(params) }
@@ -1557,7 +2095,7 @@ ${lines.join("\n")}`
     async (params) => {
       const { requireApiKey: requireApiKey2 } = await import("./config.js");
       requireApiKey2();
-      const { handleDeleteDatasetRecords } = await import("./delete-dataset-records-LTG5LT26.js");
+      const { handleDeleteDatasetRecords } = await import("./delete-dataset-records-DM4W3PAW.js");
       return {
         content: [
           { type: "text", text: await handleDeleteDatasetRecords(params) }
@@ -1565,9 +2103,144 @@ ${lines.join("\n")}`
       };
     }
   );
+  server.tool(
+    "platform_list_projects",
+    "List all projects in your LangWatch organization. Requires an org-level API key.",
+    {
+      page: z2.number().int().positive().optional().describe("Page number (default: 1)"),
+      limit: z2.number().int().positive().max(1e3).optional().describe("Results per page (default: 100)")
+    },
+    withToolLogging("platform_list_projects", async (params) => {
+      requireApiKey();
+      const { handleListProjects } = await import("./list-projects-5FHMJ4PQ.js");
+      return {
+        content: [{ type: "text", text: await handleListProjects(params) }]
+      };
+    })
+  );
+  server.tool(
+    "platform_get_project",
+    "Get details of a specific project by its ID.",
+    {
+      id: z2.string().describe("The project ID")
+    },
+    withToolLogging("platform_get_project", async (params) => {
+      requireApiKey();
+      const { handleGetProject } = await import("./get-project-JCMWXVLU.js");
+      return {
+        content: [{ type: "text", text: await handleGetProject(params) }]
+      };
+    })
+  );
+  server.tool(
+    "platform_create_project",
+    `Create a new project in your LangWatch organization. Returns a one-time service API key for the project.
+
+You must provide either teamId (to add the project to an existing team) or newTeamName (to create a new team for the project).`,
+    {
+      name: z2.string().describe("Project name"),
+      language: z2.string().describe('Programming language (e.g. "python", "typescript")'),
+      framework: z2.string().describe('Framework (e.g. "openai", "langchain", "custom")'),
+      teamId: z2.string().optional().describe("ID of an existing team to add the project to"),
+      newTeamName: z2.string().optional().describe("Name for a new team to create for this project")
+    },
+    withToolLogging("platform_create_project", async (params) => {
+      requireApiKey();
+      const { handleCreateProject } = await import("./create-project-CUVQ67VN.js");
+      return {
+        content: [{ type: "text", text: await handleCreateProject(params) }]
+      };
+    })
+  );
+  server.tool(
+    "platform_update_project",
+    "Update an existing project's name, language, framework, or PII redaction level.",
+    {
+      id: z2.string().describe("The project ID to update"),
+      name: z2.string().optional().describe("New project name"),
+      language: z2.string().optional().describe("New programming language"),
+      framework: z2.string().optional().describe("New framework"),
+      piiRedactionLevel: z2.enum(["STRICT", "ESSENTIAL", "DISABLED"]).optional().describe("PII redaction level")
+    },
+    withToolLogging("platform_update_project", async (params) => {
+      requireApiKey();
+      const { handleUpdateProject } = await import("./update-project-OIT4WPFI.js");
+      return {
+        content: [{ type: "text", text: await handleUpdateProject(params) }]
+      };
+    })
+  );
+  server.tool(
+    "platform_archive_project",
+    "Archive (soft-delete) a project. The project can be restored later if needed.",
+    {
+      id: z2.string().describe("The project ID to archive")
+    },
+    withToolLogging("platform_archive_project", async (params) => {
+      requireApiKey();
+      const { handleArchiveProject } = await import("./archive-project-HZBRPR2S.js");
+      return {
+        content: [{ type: "text", text: await handleArchiveProject(params) }]
+      };
+    })
+  );
+  server.tool(
+    "platform_list_api_keys",
+    "List all API keys in your LangWatch organization. Shows key metadata, status (active/revoked/expired), and role bindings. API key tokens are never returned.",
+    {},
+    withToolLogging("platform_list_api_keys", async () => {
+      requireApiKey();
+      const { handleListApiKeys } = await import("./list-api-keys-AWW7ROLB.js");
+      return {
+        content: [{ type: "text", text: await handleListApiKeys() }]
+      };
+    })
+  );
+  server.tool(
+    "platform_create_api_key",
+    `Create a new API key. Two types available:
+
+- **personal**: Tied to a user, requires explicit role bindings (e.g. ADMIN on PROJECT:proj_123).
+- **service**: Not tied to a user. Use projectIds for project-scoped service keys, or omit for org-wide access.
+
+The token is returned once and cannot be retrieved again.`,
+    {
+      keyType: z2.enum(["personal", "service"]).describe("Key type: 'personal' (user-tied) or 'service' (headless)"),
+      name: z2.string().describe("Key name (max 100 chars)"),
+      description: z2.string().optional().describe("Key description (max 500 chars)"),
+      expiresAt: z2.string().optional().describe("Expiration date (ISO string)"),
+      bindings: z2.array(z2.object({
+        role: z2.enum(["ADMIN", "MEMBER", "VIEWER"]).describe("Role"),
+        scopeType: z2.enum(["ORGANIZATION", "TEAM", "PROJECT"]).describe("Scope type"),
+        scopeId: z2.string().describe("Scope ID (org, team, or project ID)")
+      })).optional().describe("Role bindings (required for personal keys)"),
+      projectIds: z2.array(z2.string()).optional().describe("Project IDs for service keys (creates ADMIN bindings per project)")
+    },
+    withToolLogging("platform_create_api_key", async (params) => {
+      requireApiKey();
+      const { handleCreateApiKey } = await import("./create-api-key-OQI5PUNQ.js");
+      return {
+        content: [{ type: "text", text: await handleCreateApiKey(params) }]
+      };
+    })
+  );
+  server.tool(
+    "platform_revoke_api_key",
+    "Revoke an API key. Once revoked, the key can no longer be used for authentication.",
+    {
+      id: z2.string().describe("The API key ID to revoke")
+    },
+    withToolLogging("platform_revoke_api_key", async (params) => {
+      requireApiKey();
+      const { handleRevokeApiKey } = await import("./revoke-api-key-ZSPSPF5G.js");
+      return {
+        content: [{ type: "text", text: await handleRevokeApiKey(params) }]
+      };
+    })
+  );
 }
 
 export {
   createMcpServer
 };
-//# sourceMappingURL=chunk-B57MZ4WI.js.map
+//# sourceMappingURL=chunk-FVDAZ4XG.js.map