@langwatch/mcp-server 0.7.0 → 0.9.0

package/dist/{chunk-E5JYYU3J.js → chunk-JVLJKQJG.js}

@@ -1,6 +1,10 @@
+import {
+  LangWatchApiError,
+  makeRequest
+} from "./chunk-42ZHHHIJ.js";
 import {
   requireApiKey
-} from "./chunk-BWWY5KEF.js";
+} from "./chunk-KOO5733S.js";
 
 // src/create-mcp-server.ts
 import { McpServer } from "@modelcontextprotocol/sdk/server/mcp.js";
@@ -9,7 +13,7 @@ import { z as z2 } from "zod";
 // package.json
 var package_default = {
   name: "@langwatch/mcp-server",
-  version: "0.7.0",
+  version: "0.9.0",
   description: "An MCP server for Langwatch.",
   type: "module",
   main: "./dist/index.js",
@@ -51,19 +55,19 @@ var package_default = {
   },
   devDependencies: {
     "@ai-sdk/anthropic": "^3.0.44",
-    "@anthropic-ai/claude-code": "^2.1.72",
-    "@eslint/js": "^9.4.0",
+    "@anthropic-ai/claude-code": "^2.1.119",
+    "@eslint/js": "^10.0.1",
     "@langwatch/scenario": "^0.4.6",
     "@types/debug": "^4.1.12",
     "@types/eslint__js": "^8.42.3",
     "@types/express": "^5.0.6",
-    "@types/node": "^16.0.0",
+    "@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: "^8.57.0",
+    eslint: "^10.2.1",
     tsup: "^8.1.0",
     tsx: "^4.20.5",
     typescript: "^5.0.0",
@@ -72,7 +76,7 @@ var package_default = {
   dependencies: {
     "@modelcontextprotocol/sdk": "^1.26.0",
     "@opentelemetry/context-async-hooks": "^2.1.0",
-    "@opentelemetry/sdk-node": "^0.204.0",
+    "@opentelemetry/sdk-node": "^0.217.0",
     chalk: "^5.6.2",
     "date-fns": "^4.1.0",
     express: "^5.2.1",
@@ -84,6 +88,13 @@ var package_default = {
   },
   bin: {
     "langwatch-mcp-server": "./dist/index.js"
+  },
+  pnpm: {
+    overrides: {
+      protobufjs: ">=7.5.6 <8.0.0 || >=8.0.2",
+      "langsmith@<0.6.0": "0.6.0",
+      "fast-uri@<3.1.2": "3.1.2"
+    }
   }
 };
 
@@ -113,6 +124,379 @@ var createDatasetSchema = z.object({
   columnTypes: z.array(datasetColumnDefinitionSchema).optional().describe("Column definitions for the dataset")
 });
 
+// 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 may still be in progress, or the id may be incorrect.`,
+        "",
+        "> Try `platform_experiment_status` first to confirm the run id and that it is completed."
+      ].join("\n");
+    }
+    throw error;
+  }
+  if (!results || results.timestamps.finishedAt == null && results.timestamps.stoppedAt == null) {
+    return [
+      `# Evaluation Results: ${params.runId}`,
+      "",
+      "**Status**: results are not yet available",
+      "",
+      `Results for run \`${params.runId}\` are not ready yet.`,
+      "",
+      "> Try `platform_experiment_status` first to confirm the run id and that it is completed."
+    ].join("\n");
+  }
+  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(`**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}`);
+  }
+  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} | ${stats.errored} |`
+      );
+    }
+    lines.push("");
+  }
+  if (rows.length === 0) {
+    lines.push("_No rows matched the filter._");
+    return lines.join("\n");
+  }
+  lines.push(
+    `## Rows (${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 handleExperimentStatus(params) {
+  const status = await makeRequest(
+    "GET",
+    `/api/experiments/runs/${encodeURIComponent(params.runId)}`
+  );
+  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"'
@@ -253,7 +637,7 @@ function registerTools(server) {
     },
     withToolLogging("search_traces", async (params) => {
       requireApiKey();
-      const { handleSearchTraces } = await import("./search-traces-VYFUIMXH.js");
+      const { handleSearchTraces } = await import("./search-traces-3BBVYPC3.js");
       return {
         content: [{ type: "text", text: await handleSearchTraces(params) }]
       };
@@ -270,7 +654,7 @@ function registerTools(server) {
     },
     withToolLogging("get_trace", async (params) => {
       requireApiKey();
-      const { handleGetTrace } = await import("./get-trace-4EJAYVBG.js");
+      const { handleGetTrace } = await import("./get-trace-MRBDS3HJ.js");
       return {
         content: [{ type: "text", text: await handleGetTrace(params) }]
       };
@@ -296,7 +680,7 @@ function registerTools(server) {
     },
     withToolLogging("get_analytics", async (params) => {
       requireApiKey();
-      const { handleGetAnalytics } = await import("./get-analytics-2OHV763I.js");
+      const { handleGetAnalytics } = await import("./get-analytics-GPVJX7SF.js");
       return {
         content: [{ type: "text", text: await handleGetAnalytics(params) }]
       };
@@ -330,7 +714,7 @@ 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-IEN7O6JT.js");
+      const { handleCreatePrompt } = await import("./create-prompt-NGXKSP4X.js");
       return {
         content: [{ type: "text", text: await handleCreatePrompt(params) }]
       };
@@ -342,7 +726,7 @@ 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-VWEZ4CDE.js");
+      const { handleListPrompts } = await import("./list-prompts-VNXKHK7T.js");
       return {
         content: [{ type: "text", text: await handleListPrompts() }]
       };
@@ -366,7 +750,7 @@ NOTE: Prompts can be managed two ways. Determine which approach the user needs:
         };
       }
       requireApiKey();
-      const { handleGetPrompt } = await import("./get-prompt-WZUOLLQC.js");
+      const { handleGetPrompt } = await import("./get-prompt-5XAGWWBG.js");
       return {
         content: [{ type: "text", text: await handleGetPrompt(params) }]
       };
@@ -391,7 +775,7 @@ 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-OWQAKE7K.js");
+      const { handleUpdatePrompt } = await import("./update-prompt-ZU2DD6KM.js");
       return {
         content: [{ type: "text", text: await handleUpdatePrompt(params) }]
       };
@@ -407,7 +791,7 @@ 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-GL7EAAF3.js");
+      const { handleAssignPromptTag } = await import("./assign-prompt-tag-FKCCZPML.js");
       return {
         content: [{ type: "text", text: await handleAssignPromptTag(params) }]
       };
@@ -419,7 +803,7 @@ 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-6LY3ZPYI.js");
+      const { handleListPromptTags } = await import("./list-prompt-tags-Z2DD6FLQ.js");
       return {
         content: [{ type: "text", text: await handleListPromptTags() }]
       };
@@ -433,7 +817,7 @@ 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-CAEZ45VO.js");
+      const { handleCreatePromptTag } = await import("./create-prompt-tag-IUKYPVHH.js");
       return {
         content: [{ type: "text", text: await handleCreatePromptTag(params) }]
       };
@@ -448,7 +832,7 @@ 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-VOJJLX4M.js");
+      const { handleRenamePromptTag } = await import("./rename-prompt-tag-A4RJUK7Z.js");
       return {
         content: [{ type: "text", text: await handleRenamePromptTag(params) }]
       };
@@ -462,7 +846,7 @@ 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-ZMZUAGFW.js");
+      const { handleDeletePromptTag } = await import("./delete-prompt-tag-I6HNMSVJ.js");
       return {
         content: [{ type: "text", text: await handleDeletePromptTag(params) }]
       };
@@ -490,7 +874,7 @@ 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-6YHSMGLP.js");
+      const { handleCreateScenario } = await import("./create-scenario-FNBXALH3.js");
       return {
         content: [
           { type: "text", text: await handleCreateScenario(params) }
@@ -508,7 +892,7 @@ 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-TXD4W2ZJ.js");
+      const { handleListScenarios } = await import("./list-scenarios-CFZHXL76.js");
       return {
         content: [
           { type: "text", text: await handleListScenarios(params) }
@@ -527,7 +911,7 @@ 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-Q6LCNUYN.js");
+      const { handleGetScenario } = await import("./get-scenario-7XBCDZWX.js");
       return {
         content: [{ type: "text", text: await handleGetScenario(params) }]
       };
@@ -545,7 +929,7 @@ 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-GDFVILPN.js");
+      const { handleUpdateScenario } = await import("./update-scenario-3VUTPDRU.js");
       return {
         content: [
           { type: "text", text: await handleUpdateScenario(params) }
@@ -561,7 +945,7 @@ 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-WJ2A7ETN.js");
+      const { handleArchiveScenario } = await import("./archive-scenario-P5RRITRM.js");
       return {
         content: [
           { type: "text", text: await handleArchiveScenario(params) }
@@ -577,7 +961,7 @@ 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-FSGBX27Z.js");
+      const { handleListSuites } = await import("./list-suites-ZDOHAAXF.js");
       return {
         content: [{ type: "text", text: await handleListSuites(params) }]
       };
@@ -592,7 +976,7 @@ 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-BXK5TN7P.js");
+      const { handleGetSuite } = await import("./get-suite-4DBVHFEM.js");
       return {
         content: [{ type: "text", text: await handleGetSuite(params) }]
       };
@@ -611,7 +995,7 @@ 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-H5XOFDE5.js");
+      const { handleCreateSuite } = await import("./create-suite-DJEPD4ZB.js");
       return {
         content: [{ type: "text", text: await handleCreateSuite(params) }]
       };
@@ -631,7 +1015,7 @@ 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-BXYUYSPT.js");
+      const { handleUpdateSuite } = await import("./update-suite-C6JFDQHL.js");
       return {
         content: [{ type: "text", text: await handleUpdateSuite(params) }]
       };
@@ -645,7 +1029,7 @@ 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-BTMHGI2D.js");
+      const { handleRunSuite } = await import("./run-suite-TH7LO5TJ.js");
       return {
         content: [{ type: "text", text: await handleRunSuite(params) }]
       };
@@ -659,7 +1043,7 @@ 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-4JOM4SSL.js");
+      const { handleArchiveSuite } = await import("./archive-suite-NUUFD2EM.js");
       return {
         content: [{ type: "text", text: await handleArchiveSuite(params) }]
       };
@@ -676,7 +1060,7 @@ 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-RHPOLMFI.js");
+      const { handleListSimulationRuns } = await import("./list-simulation-runs-AM2BXLBW.js");
       return {
         content: [{ type: "text", text: await handleListSimulationRuns(params) }]
       };
@@ -691,7 +1075,7 @@ 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-Z2XKO256.js");
+      const { handleGetSimulationRun } = await import("./get-simulation-run-AV24FJWI.js");
       return {
         content: [{ type: "text", text: await handleGetSimulationRun(params) }]
       };
@@ -708,7 +1092,7 @@ 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-WYFRFLRC.js");
+      const { handleCreateEvaluator } = await import("./create-evaluator-YFFQIR7E.js");
       return {
         content: [
           { type: "text", text: await handleCreateEvaluator(params) }
@@ -722,7 +1106,7 @@ 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-7R2LB4JH.js");
+      const { handleListEvaluators } = await import("./list-evaluators-B6OSVAM2.js");
       return {
         content: [{ type: "text", text: await handleListEvaluators() }]
       };
@@ -736,7 +1120,7 @@ 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-UWVUPYTM.js");
+      const { handleGetEvaluator } = await import("./get-evaluator-PCM6XQAU.js");
       return {
         content: [{ type: "text", text: await handleGetEvaluator(params) }]
       };
@@ -754,7 +1138,7 @@ 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-IVOR6FTR.js");
+      const { handleUpdateEvaluator } = await import("./update-evaluator-6IJHFG2E.js");
       return {
         content: [
           { type: "text", text: await handleUpdateEvaluator(params) }
@@ -770,7 +1154,7 @@ 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-7FVC5WM4.js");
+      const { handleDeleteEvaluator } = await import("./delete-evaluator-7ZATLJOA.js");
       return {
         content: [{ type: "text", text: await handleDeleteEvaluator(params) }]
       };
@@ -791,7 +1175,7 @@ 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-W5IUI326.js");
+      const { handleSetModelProvider } = await import("./set-model-provider-WOC5FNMV.js");
       return {
         content: [
           { type: "text", text: await handleSetModelProvider(params) }
@@ -805,7 +1189,7 @@ 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-OBYJCKBX.js");
+      const { handleListModelProviders } = await import("./list-model-providers-NL6IP4UZ.js");
       return {
         content: [{ type: "text", text: await handleListModelProviders() }]
       };
@@ -817,7 +1201,7 @@ 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-AW3P6FJV.js");
+      const { handleListAgents } = await import("./list-agents-NECO7MPI.js");
       return {
         content: [{ type: "text", text: await handleListAgents() }]
       };
@@ -831,7 +1215,7 @@ 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-E3QSTVKG.js");
+      const { handleGetAgent } = await import("./get-agent-3KDWCP5N.js");
       return {
         content: [{ type: "text", text: await handleGetAgent(params) }]
       };
@@ -847,7 +1231,7 @@ 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-MBOOT3XK.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 }) }]
@@ -865,7 +1249,7 @@ 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-4BB7EYLR.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 }) }]
@@ -880,7 +1264,7 @@ 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-JKTNBWAA.js");
+      const { handleDeleteAgent } = await import("./delete-agent-TJI5XBBM.js");
       return {
         content: [{ type: "text", text: await handleDeleteAgent(params) }]
       };
@@ -895,7 +1279,7 @@ 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-KTCC7TB7.js");
+      const { handleRunAgent } = await import("./run-agent-JFNDY7MR.js");
       return {
         content: [{ type: "text", text: await handleRunAgent(params) }]
       };
@@ -907,7 +1291,7 @@ 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-R7RK2IXL.js");
+      const { handleListDashboards } = await import("./list-dashboards-DVPVVVKW.js");
       return {
         content: [{ type: "text", text: await handleListDashboards() }]
       };
@@ -921,7 +1305,7 @@ 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-PC7JDHH6.js");
+      const { handleGetDashboard } = await import("./get-dashboard-YPTIE5YY.js");
       return {
         content: [{ type: "text", text: await handleGetDashboard(params) }]
       };
@@ -935,7 +1319,7 @@ 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-VRKLFVZJ.js");
+      const { handleCreateDashboard } = await import("./create-dashboard-G4QA5AXP.js");
       return {
         content: [{ type: "text", text: await handleCreateDashboard(params) }]
       };
@@ -949,7 +1333,7 @@ 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-VJSR7CRU.js");
+      const { handleDeleteDashboard } = await import("./delete-dashboard-N5SZJIR3.js");
       return {
         content: [{ type: "text", text: await handleDeleteDashboard(params) }]
       };
@@ -964,7 +1348,7 @@ 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-G6TR7VVO.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}).` }]
@@ -977,7 +1361,7 @@ 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-VRGEZ3EN.js");
+      const { handleListWorkflows } = await import("./list-workflows-WBRBKLIT.js");
       return {
         content: [{ type: "text", text: await handleListWorkflows() }]
       };
@@ -991,7 +1375,7 @@ 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-O34QJLHD.js");
+      const { handleGetWorkflow } = await import("./get-workflow-UR4Y6UFX.js");
       return {
         content: [{ type: "text", text: await handleGetWorkflow(params) }]
       };
@@ -1005,7 +1389,7 @@ 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-EYQ3T4RQ.js");
+      const { handleDeleteWorkflow } = await import("./delete-workflow-QEO74JQR.js");
       return {
         content: [{ type: "text", text: await handleDeleteWorkflow(params) }]
       };
@@ -1020,7 +1404,7 @@ 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-IGAYGKJL.js");
+      const { handleRunWorkflow } = await import("./run-workflow-CK3JP5U2.js");
       return {
         content: [{ type: "text", text: await handleRunWorkflow(params) }]
       };
@@ -1034,7 +1418,7 @@ 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-QNCGWPQ3.js");
+      const { handleListAnnotations } = await import("./list-annotations-3OJEO774.js");
       return {
         content: [{ type: "text", text: await handleListAnnotations(params) }]
       };
@@ -1051,7 +1435,7 @@ 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-XMV6A7ED.js");
+      const { handleCreateAnnotation } = await import("./create-annotation-KN3GE3DK.js");
       return {
         content: [{ type: "text", text: await handleCreateAnnotation(params) }]
       };
@@ -1065,7 +1449,7 @@ 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-ACMDV6QO.js");
+      const { handleDeleteAnnotation } = await import("./delete-annotation-KRP53TUU.js");
       return {
         content: [{ type: "text", text: await handleDeleteAnnotation(params) }]
       };
@@ -1079,7 +1463,7 @@ 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-ERBG27RF.js");
+      const { getAnnotation } = await import("./langwatch-api-annotations-ZJDRB4JH.js");
       const annotation = await getAnnotation(params.id);
       const lines = [
         `**ID**: ${annotation.id}`,
@@ -1101,7 +1485,7 @@ NOTE: Scenarios can be created two ways. Determine which approach the user needs
     },
     withToolLogging("platform_list_triggers", async (params) => {
       requireApiKey();
-      const { listTriggers } = await import("./langwatch-api-triggers-2RTNSWQY.js");
+      const { listTriggers } = await import("./langwatch-api-triggers-5SX6R2WU.js");
       const triggers = await listTriggers();
       if (params.format === "json") {
         return { content: [{ type: "text", text: JSON.stringify(triggers, null, 2) }] };
@@ -1134,7 +1518,7 @@ NOTE: Scenarios can be created two ways. Determine which approach the user needs
     },
     withToolLogging("platform_create_trigger", async (params) => {
       requireApiKey();
-      const { createTrigger } = await import("./langwatch-api-triggers-2RTNSWQY.js");
+      const { createTrigger } = await import("./langwatch-api-triggers-5SX6R2WU.js");
       let filters = {};
       if (params.filters) {
         try {
@@ -1165,7 +1549,7 @@ NOTE: Scenarios can be created two ways. Determine which approach the user needs
     },
     withToolLogging("platform_update_trigger", async (params) => {
       requireApiKey();
-      const { updateTrigger } = await import("./langwatch-api-triggers-2RTNSWQY.js");
+      const { updateTrigger } = await import("./langwatch-api-triggers-5SX6R2WU.js");
       const trigger = await updateTrigger(params);
       return { content: [{ type: "text", text: `Trigger "${trigger.name}" updated (active: ${trigger.active}).` }] };
     })
@@ -1178,7 +1562,7 @@ 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-2RTNSWQY.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.` }] };
     })
@@ -1189,7 +1573,7 @@ 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-WSXZLS2B.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." }] };
@@ -1216,7 +1600,7 @@ ${lines.join("\n")}`
     },
     withToolLogging("platform_get_monitor", async (params) => {
       requireApiKey();
-      const { getMonitor } = await import("./langwatch-api-monitors-WSXZLS2B.js");
+      const { getMonitor } = await import("./langwatch-api-monitors-4XKXHYCG.js");
       const m = await getMonitor(params.id);
       const lines = [
         `**${m.name}** (${m.id})`,
@@ -1243,7 +1627,7 @@ ${lines.join("\n")}`
     },
     withToolLogging("platform_create_monitor", async (params) => {
       requireApiKey();
-      const { createMonitor } = await import("./langwatch-api-monitors-WSXZLS2B.js");
+      const { createMonitor } = await import("./langwatch-api-monitors-4XKXHYCG.js");
       const monitor = await createMonitor(params);
       return {
         content: [
@@ -1267,7 +1651,7 @@ ${lines.join("\n")}`
     },
     withToolLogging("platform_update_monitor", async (params) => {
       requireApiKey();
-      const { updateMonitor } = await import("./langwatch-api-monitors-WSXZLS2B.js");
+      const { updateMonitor } = await import("./langwatch-api-monitors-4XKXHYCG.js");
       const monitor = await updateMonitor(params);
       return {
         content: [
@@ -1287,7 +1671,7 @@ ${lines.join("\n")}`
     },
     withToolLogging("platform_delete_monitor", async (params) => {
       requireApiKey();
-      const { deleteMonitor } = await import("./langwatch-api-monitors-WSXZLS2B.js");
+      const { deleteMonitor } = await import("./langwatch-api-monitors-4XKXHYCG.js");
       const result = await deleteMonitor(params.id);
       return {
         content: [
@@ -1302,7 +1686,7 @@ ${lines.join("\n")}`
     {},
     withToolLogging("platform_list_secrets", async () => {
       requireApiKey();
-      const { listSecrets } = await import("./langwatch-api-secrets-RR5DKG2X.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." }] };
@@ -1330,7 +1714,7 @@ ${lines.join("\n")}`
     },
     withToolLogging("platform_create_secret", async (params) => {
       requireApiKey();
-      const { createSecret } = await import("./langwatch-api-secrets-RR5DKG2X.js");
+      const { createSecret } = await import("./langwatch-api-secrets-6SHK5G7Q.js");
       const secret = await createSecret({ name: params.name, value: params.value });
       return {
         content: [
@@ -1351,7 +1735,7 @@ ${lines.join("\n")}`
     },
     withToolLogging("platform_update_secret", async (params) => {
       requireApiKey();
-      const { updateSecret } = await import("./langwatch-api-secrets-RR5DKG2X.js");
+      const { updateSecret } = await import("./langwatch-api-secrets-6SHK5G7Q.js");
       const secret = await updateSecret({ id: params.id, value: params.value });
       return {
         content: [
@@ -1371,36 +1755,92 @@ ${lines.join("\n")}`
     },
     withToolLogging("platform_delete_secret", async (params) => {
       requireApiKey();
-      const { deleteSecret } = await import("./langwatch-api-secrets-RR5DKG2X.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. Returns progress and summary when completed.",
+    {
+      runId: z2.string().describe("The run ID returned from platform_run_experiment")
+    },
+    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.",
     {
-      slug: z2.string().describe("The evaluation slug to run")
+      limit: z2.number().int().positive().max(100).optional().describe(
+        "Maximum number of experiments to include (default 25, hard-capped at 100)"
+      )
     },
-    withToolLogging("platform_run_evaluation", async (params) => {
+    withToolLogging("platform_experiment_list", async (params) => {
       requireApiKey();
-      const { handleRunEvaluation } = await import("./run-evaluation-2X3V3Q6L.js");
       return {
-        content: [{ type: "text", text: await handleRunEvaluation(params) }]
+        content: [{ type: "text", text: await handleExperimentList(params) }]
       };
     })
   );
   server.tool(
-    "platform_evaluation_status",
-    "Check the status of an evaluation run. Returns progress and summary when completed.",
+    "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.",
     {
-      runId: z2.string().describe("The run ID returned from platform_run_evaluation")
+      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_evaluation_status", async (params) => {
+    withToolLogging("platform_experiment_list_runs", async (params) => {
       requireApiKey();
-      const { handleEvaluationStatus } = await import("./run-evaluation-2X3V3Q6L.js");
       return {
-        content: [{ type: "text", text: await handleEvaluationStatus(params) }]
+        content: [{ type: "text", text: await handleExperimentListRuns(params) }]
+      };
+    })
+  );
+  server.tool(
+    "platform_experiment_results",
+    "Fetch per-row results for a completed experiment run so you can debug evaluator scores and missed rows. 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_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_experiment_results", async (params) => {
+      requireApiKey();
+      return {
+        content: [
+          { type: "text", text: await handleExperimentResults(params) }
+        ]
       };
     })
   );
@@ -1415,7 +1855,7 @@ ${lines.join("\n")}`
     async (params) => {
       const { requireApiKey: requireApiKey2 } = await import("./config.js");
       requireApiKey2();
-      const { handleListDatasets } = await import("./list-datasets-KEXRLDEU.js");
+      const { handleListDatasets } = await import("./list-datasets-WUAE5HGX.js");
       return {
         content: [{ type: "text", text: await handleListDatasets(params) }]
       };
@@ -1433,7 +1873,7 @@ ${lines.join("\n")}`
     async (params) => {
       const { requireApiKey: requireApiKey2 } = await import("./config.js");
       requireApiKey2();
-      const { handleGetDataset } = await import("./get-dataset-H5IK5FVU.js");
+      const { handleGetDataset } = await import("./get-dataset-6DQFYITM.js");
       return {
         content: [{ type: "text", text: await handleGetDataset(params) }]
       };
@@ -1446,7 +1886,7 @@ ${lines.join("\n")}`
     async (params) => {
       const { requireApiKey: requireApiKey2 } = await import("./config.js");
       requireApiKey2();
-      const { handleCreateDataset } = await import("./create-dataset-7HQPZCRE.js");
+      const { handleCreateDataset } = await import("./create-dataset-WQSC37N7.js");
       return {
         content: [{ type: "text", text: await handleCreateDataset(params) }]
       };
@@ -1463,7 +1903,7 @@ ${lines.join("\n")}`
     async (params) => {
       const { requireApiKey: requireApiKey2 } = await import("./config.js");
       requireApiKey2();
-      const { handleUpdateDataset } = await import("./update-dataset-CXBKZ4W3.js");
+      const { handleUpdateDataset } = await import("./update-dataset-EYT5POFM.js");
       return {
         content: [{ type: "text", text: await handleUpdateDataset(params) }]
       };
@@ -1478,7 +1918,7 @@ ${lines.join("\n")}`
     async (params) => {
       const { requireApiKey: requireApiKey2 } = await import("./config.js");
       requireApiKey2();
-      const { handleDeleteDataset } = await import("./delete-dataset-RTTTWP4D.js");
+      const { handleDeleteDataset } = await import("./delete-dataset-KDUOHL2P.js");
       return {
         content: [{ type: "text", text: await handleDeleteDataset(params) }]
       };
@@ -1496,7 +1936,7 @@ ${lines.join("\n")}`
     async (params) => {
       const { requireApiKey: requireApiKey2 } = await import("./config.js");
       requireApiKey2();
-      const { handleListDatasetRecords } = await import("./list-dataset-records-WVCSQP7P.js");
+      const { handleListDatasetRecords } = await import("./list-dataset-records-I4BYORF5.js");
       return {
         content: [
           { type: "text", text: await handleListDatasetRecords(params) }
@@ -1514,7 +1954,7 @@ ${lines.join("\n")}`
     async (params) => {
       const { requireApiKey: requireApiKey2 } = await import("./config.js");
       requireApiKey2();
-      const { handleCreateDatasetRecords } = await import("./create-dataset-records-2NIVFYTV.js");
+      const { handleCreateDatasetRecords } = await import("./create-dataset-records-B4RL3FLC.js");
       return {
         content: [
           { type: "text", text: await handleCreateDatasetRecords(params) }
@@ -1533,7 +1973,7 @@ ${lines.join("\n")}`
     async (params) => {
       const { requireApiKey: requireApiKey2 } = await import("./config.js");
       requireApiKey2();
-      const { handleUpdateDatasetRecord } = await import("./update-dataset-record-HOHAMHXG.js");
+      const { handleUpdateDatasetRecord } = await import("./update-dataset-record-VFF5JTX4.js");
       return {
         content: [
           { type: "text", text: await handleUpdateDatasetRecord(params) }
@@ -1551,7 +1991,7 @@ ${lines.join("\n")}`
     async (params) => {
       const { requireApiKey: requireApiKey2 } = await import("./config.js");
       requireApiKey2();
-      const { handleDeleteDatasetRecords } = await import("./delete-dataset-records-KNQTIPQK.js");
+      const { handleDeleteDatasetRecords } = await import("./delete-dataset-records-DM4W3PAW.js");
       return {
         content: [
           { type: "text", text: await handleDeleteDatasetRecords(params) }
@@ -1559,9 +1999,144 @@ ${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-E5JYYU3J.js.map
+//# sourceMappingURL=chunk-JVLJKQJG.js.map