@langwatch/mcp-server 1.0.0 → 2.1.0

package/dist/{discover-evaluator-schema-MRGHVNTT.js → discover-evaluator-schema-ESSHQGPY.js}

@@ -1,6 +1,6 @@
-import "./chunk-BJ7B4R26.js";
+import "./chunk-SAS5OB7F.js";
 
-// ../langevals/ts-integration/evaluators.generated.ts
+// ../../services/langevals/ts-integration/evaluators.generated.ts
 import { z } from "zod";
 var evaluatorTypesSchema = z.string();
 var moneySchema = z.object({
@@ -18,7 +18,8 @@ var evaluationResultSchema = z.object({
 });
 var evaluationResultSkippedSchema = z.object({
   status: z.literal("skipped"),
-  details: z.string().optional()
+  details: z.string().optional(),
+  cost: moneySchema.optional()
 });
 var evaluationResultErrorSchema = z.object({
   status: z.literal("error"),
@@ -379,69 +380,6 @@ var evaluatorsSchema = z.object({
       ).default(0.25)
     })
   }),
-  "legacy/ragas_answer_correctness": z.object({
-    settings: z.object({
-      model: z.string().describe("The model to use for evaluation.").default("openai/gpt-5-mini"),
-      embeddings_model: z.string().describe("The model to use for embeddings.").default("openai/text-embedding-ada-002"),
-      max_tokens: z.number().describe(
-        "The maximum number of tokens allowed for evaluation, a too high number can be costly. Entries above this amount will be skipped."
-      ).default(2048)
-    })
-  }),
-  "legacy/ragas_answer_relevancy": z.object({
-    settings: z.object({
-      model: z.string().describe("The model to use for evaluation.").default("openai/gpt-5-mini"),
-      embeddings_model: z.string().describe("The model to use for embeddings.").default("openai/text-embedding-ada-002"),
-      max_tokens: z.number().describe(
-        "The maximum number of tokens allowed for evaluation, a too high number can be costly. Entries above this amount will be skipped."
-      ).default(2048)
-    })
-  }),
-  "legacy/ragas_context_precision": z.object({
-    settings: z.object({
-      model: z.string().describe("The model to use for evaluation.").default("openai/gpt-5-mini"),
-      embeddings_model: z.string().describe("The model to use for embeddings.").default("openai/text-embedding-ada-002"),
-      max_tokens: z.number().describe(
-        "The maximum number of tokens allowed for evaluation, a too high number can be costly. Entries above this amount will be skipped."
-      ).default(2048)
-    })
-  }),
-  "legacy/ragas_context_recall": z.object({
-    settings: z.object({
-      model: z.string().describe("The model to use for evaluation.").default("openai/gpt-5-mini"),
-      embeddings_model: z.string().describe("The model to use for embeddings.").default("openai/text-embedding-ada-002"),
-      max_tokens: z.number().describe(
-        "The maximum number of tokens allowed for evaluation, a too high number can be costly. Entries above this amount will be skipped."
-      ).default(2048)
-    })
-  }),
-  "legacy/ragas_context_relevancy": z.object({
-    settings: z.object({
-      model: z.string().describe("The model to use for evaluation.").default("openai/gpt-5-mini"),
-      embeddings_model: z.string().describe("The model to use for embeddings.").default("openai/text-embedding-ada-002"),
-      max_tokens: z.number().describe(
-        "The maximum number of tokens allowed for evaluation, a too high number can be costly. Entries above this amount will be skipped."
-      ).default(2048)
-    })
-  }),
-  "legacy/ragas_context_utilization": z.object({
-    settings: z.object({
-      model: z.string().describe("The model to use for evaluation.").default("openai/gpt-5-mini"),
-      embeddings_model: z.string().describe("The model to use for embeddings.").default("openai/text-embedding-ada-002"),
-      max_tokens: z.number().describe(
-        "The maximum number of tokens allowed for evaluation, a too high number can be costly. Entries above this amount will be skipped."
-      ).default(2048)
-    })
-  }),
-  "legacy/ragas_faithfulness": z.object({
-    settings: z.object({
-      model: z.string().describe("The model to use for evaluation.").default("openai/gpt-5-mini"),
-      embeddings_model: z.string().describe("The model to use for embeddings.").default("openai/text-embedding-ada-002"),
-      max_tokens: z.number().describe(
-        "The maximum number of tokens allowed for evaluation, a too high number can be costly. Entries above this amount will be skipped."
-      ).default(2048)
-    })
-  }),
   "langevals/basic": z.object({
     settings: z.object({
       rules: z.array(
@@ -580,6 +518,12 @@ var evaluatorsSchema = z.object({
       include_metrics: z.array(z.union([z.literal("cost"), z.literal("duration")])).describe("Per-candidate metrics to inject into the judge prompt").default([])
     })
   }),
+  "langevals/query_resolution": z.object({
+    settings: z.object({
+      model: z.string().describe("The model to use for evaluation").default("openai/gpt-5-mini"),
+      max_tokens: z.number().describe("Max tokens allowed for evaluation").default(128e3)
+    })
+  }),
   "langevals/select_best_compare": z.object({
     settings: z.object({
       model: z.string().describe("The model to use for evaluation").default("openai/gpt-5-mini"),
@@ -609,13 +553,13 @@ slot label. Use "tie" only when no candidate is clearly better.
       allow_tie: z.boolean().describe(
         "Allow the judge to return 'tie' when candidates are equivalent"
       ).default(true),
-      include_metrics: z.array(z.union([z.literal("cost"), z.literal("duration")])).describe("Per-candidate metrics to inject into the judge prompt").default([])
-    })
-  }),
-  "langevals/query_resolution": z.object({
-    settings: z.object({
-      model: z.string().describe("The model to use for evaluation").default("openai/gpt-5-mini"),
-      max_tokens: z.number().describe("Max tokens allowed for evaluation").default(128e3)
+      include_metrics: z.array(z.union([z.literal("cost"), z.literal("duration")])).describe("Per-candidate metrics to inject into the judge prompt").default([]),
+      temperature: z.number().describe(
+        "Sampling temperature for the judge call. Lower is more deterministic."
+      ).default(0),
+      swap_and_reconcile: z.boolean().describe(
+        "Check each row twice, the second time with the candidate order reversed, and record no result for that row when the two disagree \u2014 rather than a tie, which would claim the candidates are equally good. Doubles judge-call cost per row (literature: swapping candidate order alone flips 10-30% of verdicts on contested rows)."
+      ).default(true)
     })
   }),
   "langevals/sentiment": z.object({
@@ -1237,223 +1181,6 @@ or if it's in a specific expected language.
       }
     }
   },
-  "legacy/ragas_answer_correctness": {
-    name: `Ragas Answer Correctness`,
-    description: `
-Computes with an LLM a weighted combination of factual as well as semantic similarity between the generated answer and the expected output.
-`,
-    category: "rag",
-    docsUrl: "https://docs.ragas.io/en/latest/concepts/metrics/answer_correctness.html",
-    isGuardrail: false,
-    requiredFields: ["output", "expected_output"],
-    optionalFields: ["input"],
-    settings: {
-      model: {
-        description: "The model to use for evaluation.",
-        default: "openai/gpt-5-mini"
-      },
-      embeddings_model: {
-        description: "The model to use for embeddings.",
-        default: "openai/text-embedding-ada-002"
-      },
-      max_tokens: {
-        description: "The maximum number of tokens allowed for evaluation, a too high number can be costly. Entries above this amount will be skipped.",
-        default: 2048
-      }
-    },
-    envVars: [],
-    result: {
-      score: {
-        description: "A score between 0.0 and 1.0 indicating the correctness of the answer."
-      }
-    }
-  },
-  "legacy/ragas_answer_relevancy": {
-    name: `Ragas Answer Relevancy`,
-    description: `
-Evaluates how pertinent the generated answer is to the given prompt. Higher scores indicate better relevancy.
-`,
-    category: "rag",
-    docsUrl: "https://docs.ragas.io/en/latest/concepts/metrics/answer_relevance.html",
-    isGuardrail: false,
-    requiredFields: ["input", "output"],
-    optionalFields: [],
-    settings: {
-      model: {
-        description: "The model to use for evaluation.",
-        default: "openai/gpt-5-mini"
-      },
-      embeddings_model: {
-        description: "The model to use for embeddings.",
-        default: "openai/text-embedding-ada-002"
-      },
-      max_tokens: {
-        description: "The maximum number of tokens allowed for evaluation, a too high number can be costly. Entries above this amount will be skipped.",
-        default: 2048
-      }
-    },
-    envVars: [],
-    result: {
-      score: {
-        description: "A score between 0.0 and 1.0 indicating the relevance of the answer."
-      }
-    }
-  },
-  "legacy/ragas_context_precision": {
-    name: `Ragas Context Precision`,
-    description: `
-This metric evaluates whether all of the ground-truth relevant items present in the contexts are ranked higher or not. Higher scores indicate better precision.
-`,
-    category: "rag",
-    docsUrl: "https://docs.ragas.io/en/latest/concepts/metrics/context_precision.html",
-    isGuardrail: false,
-    requiredFields: ["input", "contexts", "expected_output"],
-    optionalFields: [],
-    settings: {
-      model: {
-        description: "The model to use for evaluation.",
-        default: "openai/gpt-5-mini"
-      },
-      embeddings_model: {
-        description: "The model to use for embeddings.",
-        default: "openai/text-embedding-ada-002"
-      },
-      max_tokens: {
-        description: "The maximum number of tokens allowed for evaluation, a too high number can be costly. Entries above this amount will be skipped.",
-        default: 2048
-      }
-    },
-    envVars: [],
-    result: {
-      score: {
-        description: "A score between 0.0 and 1.0 indicating the precision of the context."
-      }
-    }
-  },
-  "legacy/ragas_context_recall": {
-    name: `Ragas Context Recall`,
-    description: `
-This evaluator measures the extent to which the retrieved context aligns with the annotated answer, treated as the ground truth. Higher values indicate better performance.
-`,
-    category: "rag",
-    docsUrl: "https://docs.ragas.io/en/latest/concepts/metrics/context_recall.html",
-    isGuardrail: false,
-    requiredFields: ["input", "contexts", "expected_output"],
-    optionalFields: [],
-    settings: {
-      model: {
-        description: "The model to use for evaluation.",
-        default: "openai/gpt-5-mini"
-      },
-      embeddings_model: {
-        description: "The model to use for embeddings.",
-        default: "openai/text-embedding-ada-002"
-      },
-      max_tokens: {
-        description: "The maximum number of tokens allowed for evaluation, a too high number can be costly. Entries above this amount will be skipped.",
-        default: 2048
-      }
-    },
-    envVars: [],
-    result: {
-      score: {
-        description: "A score between 0.0 and 1.0 indicating the recall of the context."
-      }
-    }
-  },
-  "legacy/ragas_context_relevancy": {
-    name: `Ragas Context Relevancy`,
-    description: `
-This metric gauges the relevancy of the retrieved context, calculated based on both the question and contexts. The values fall within the range of (0, 1), with higher values indicating better relevancy.
-`,
-    category: "rag",
-    docsUrl: "https://docs.ragas.io/en/latest/concepts/metrics/context_relevancy.html",
-    isGuardrail: false,
-    requiredFields: ["output", "contexts"],
-    optionalFields: [],
-    settings: {
-      model: {
-        description: "The model to use for evaluation.",
-        default: "openai/gpt-5-mini"
-      },
-      embeddings_model: {
-        description: "The model to use for embeddings.",
-        default: "openai/text-embedding-ada-002"
-      },
-      max_tokens: {
-        description: "The maximum number of tokens allowed for evaluation, a too high number can be costly. Entries above this amount will be skipped.",
-        default: 2048
-      }
-    },
-    envVars: [],
-    result: {
-      score: {
-        description: "A score between 0.0 and 1.0 indicating the relevancy of the context."
-      }
-    }
-  },
-  "legacy/ragas_context_utilization": {
-    name: `Ragas Context Utilization`,
-    description: `
-This metric evaluates whether all of the output relevant items present in the contexts are ranked higher or not. Higher scores indicate better utilization.
-`,
-    category: "rag",
-    docsUrl: "https://docs.ragas.io/en/latest/concepts/metrics/context_precision.html",
-    isGuardrail: false,
-    requiredFields: ["input", "output", "contexts"],
-    optionalFields: [],
-    settings: {
-      model: {
-        description: "The model to use for evaluation.",
-        default: "openai/gpt-5-mini"
-      },
-      embeddings_model: {
-        description: "The model to use for embeddings.",
-        default: "openai/text-embedding-ada-002"
-      },
-      max_tokens: {
-        description: "The maximum number of tokens allowed for evaluation, a too high number can be costly. Entries above this amount will be skipped.",
-        default: 2048
-      }
-    },
-    envVars: [],
-    result: {
-      score: {
-        description: "A score between 0.0 and 1.0 indicating the utilization of the context."
-      }
-    }
-  },
-  "legacy/ragas_faithfulness": {
-    name: `Ragas Faithfulness`,
-    description: `
-This evaluator assesses the extent to which the generated answer is consistent with the provided context. Higher scores indicate better faithfulness to the context, useful for detecting hallucinations.
-`,
-    category: "rag",
-    docsUrl: "https://docs.ragas.io/en/latest/concepts/metrics/faithfulness.html",
-    isGuardrail: false,
-    requiredFields: ["output", "contexts"],
-    optionalFields: [],
-    settings: {
-      model: {
-        description: "The model to use for evaluation.",
-        default: "openai/gpt-5-mini"
-      },
-      embeddings_model: {
-        description: "The model to use for embeddings.",
-        default: "openai/text-embedding-ada-002"
-      },
-      max_tokens: {
-        description: "The maximum number of tokens allowed for evaluation, a too high number can be costly. Entries above this amount will be skipped.",
-        default: 2048
-      }
-    },
-    envVars: [],
-    result: {
-      score: {
-        description: "A score between 0.0 and 1.0 indicating the faithfulness of the answer."
-      }
-    }
-  },
   "langevals/basic": {
     name: `Custom Basic Evaluator`,
     description: `
@@ -1597,8 +1324,8 @@ A simple evaluator that checks if the output matches the expected_output exactly
     category: "quality",
     docsUrl: "",
     isGuardrail: false,
-    requiredFields: [],
-    optionalFields: ["output", "expected_output"],
+    requiredFields: ["output", "expected_output"],
+    optionalFields: [],
     settings: {
       case_sensitive: {
         description: "True if the comparison should be case-sensitive, False otherwise",
@@ -1628,8 +1355,8 @@ Uses an LLM to check if the generated output answers a question correctly the sa
     category: "quality",
     docsUrl: "",
     isGuardrail: false,
-    requiredFields: [],
-    optionalFields: ["input", "output", "expected_output"],
+    requiredFields: ["output", "expected_output"],
+    optionalFields: ["input"],
     settings: {
       model: {
         description: "The model to use for evaluation",
@@ -1866,19 +1593,44 @@ better one. Superseded by Comparison, which judges two or more candidates.
       }
     }
   },
+  "langevals/query_resolution": {
+    name: `Query Resolution`,
+    description: `
+This evaluator checks if all the user queries in the conversation were resolved. Useful to detect when the bot doesn't know how to answer or can't help the user.
+`,
+    category: "quality",
+    docsUrl: "",
+    isGuardrail: false,
+    requiredFields: ["conversation"],
+    optionalFields: [],
+    settings: {
+      model: {
+        description: "The model to use for evaluation",
+        default: "openai/gpt-5-mini"
+      },
+      max_tokens: {
+        description: "Max tokens allowed for evaluation",
+        default: 128e3
+      }
+    },
+    envVars: [],
+    result: {}
+  },
   "langevals/select_best_compare": {
     name: `Comparison`,
     description: `
 Compare two or more candidate outputs and pick the best one, optionally
 against a reference answer. The judge sees every candidate at once and
 explains why the winner is better. Candidate order is shuffled so that a
-candidate's position never sways the verdict.
+candidate's position never sways the verdict, and by default each row is
+checked twice \u2014 once more with the order reversed \u2014 recording no result
+for that row when the two checks disagree, rather than guessing.
 `,
     category: "quality",
     docsUrl: "",
     isGuardrail: false,
-    requiredFields: ["candidates"],
-    optionalFields: ["input", "golden", "row_index"],
+    requiredFields: [],
+    optionalFields: ["input", "golden", "candidates", "row_index"],
     settings: {
       model: {
         description: "The model to use for evaluation",
@@ -1918,6 +1670,14 @@ slot label. Use "tie" only when no candidate is clearly better.
       include_metrics: {
         description: "Per-candidate metrics to inject into the judge prompt",
         default: []
+      },
+      temperature: {
+        description: "Sampling temperature for the judge call. Lower is more deterministic.",
+        default: 0
+      },
+      swap_and_reconcile: {
+        description: "Check each row twice, the second time with the candidate order reversed, and record no result for that row when the two disagree \u2014 rather than a tie, which would claim the candidates are equally good. Doubles judge-call cost per row (literature: swapping candidate order alone flips 10-30% of verdicts on contested rows).",
+        default: true
       }
     },
     envVars: [],
@@ -1930,29 +1690,6 @@ slot label. Use "tie" only when no candidate is clearly better.
       }
     }
   },
-  "langevals/query_resolution": {
-    name: `Query Resolution`,
-    description: `
-This evaluator checks if all the user queries in the conversation were resolved. Useful to detect when the bot doesn't know how to answer or can't help the user.
-`,
-    category: "quality",
-    docsUrl: "",
-    isGuardrail: false,
-    requiredFields: ["conversation"],
-    optionalFields: [],
-    settings: {
-      model: {
-        description: "The model to use for evaluation",
-        default: "openai/gpt-5-mini"
-      },
-      max_tokens: {
-        description: "Max tokens allowed for evaluation",
-        default: 128e3
-      }
-    },
-    envVars: [],
-    result: {}
-  },
   "langevals/sentiment": {
     name: `Sentiment Evaluator`,
     description: `
@@ -2172,4 +1909,4 @@ function extractFirstLine(description) {
 export {
   formatEvaluatorSchema
 };
-//# sourceMappingURL=discover-evaluator-schema-MRGHVNTT.js.map
+//# sourceMappingURL=discover-evaluator-schema-ESSHQGPY.js.map

package/dist/{discover-scenario-schema-UPBTCWPG.js → discover-scenario-schema-W7KBXA4J.js}

@@ -1,4 +1,4 @@
-import "./chunk-BJ7B4R26.js";
+import "./chunk-SAS5OB7F.js";
 
 // src/tools/discover-scenario-schema.ts
 function formatScenarioSchema() {
@@ -17,6 +17,58 @@ function formatScenarioSchema() {
   lines.push(
     '- **labels** (array of strings): Tags for organizing scenarios (e.g., "auth", "happy-path", "edge-case")'
   );
+  lines.push(
+    "- **testSuiteId** (string): The test suite the scenario is filed in. Every project has a Default suite."
+  );
+  lines.push(
+    `- **fields** (object): The scenario's value for each field its test suite declares, by identifier (e.g. { "golden_sql": "SELECT ..." }). A text field takes a string, a number field a number, a boolean field true or false. A field the suite does not declare is refused. A blank value skips the evaluators that read it.`
+  );
+  lines.push("\n## Test Suite Fields\n");
+  lines.push(
+    "A test suite reads like a dataset: it declares typed fields beyond the situation and the criteria, and every scenario filed in it carries one value per field. Declare them with `platform_create_test_suite` or `platform_update_test_suite`:"
+  );
+  lines.push("```json");
+  lines.push('{ "fields": [{ "identifier": "golden_sql", "type": "text" }, { "identifier": "row_limit", "type": "number" }] }');
+  lines.push("```");
+  lines.push(
+    "- **identifier**: lowercase letters, digits and underscores, starting with a letter. situation, criteria, name, input and output are reserved."
+  );
+  lines.push("- **type**: text, number or boolean.");
+  lines.push(
+    "- On update, the list replaces the one the suite holds. A field an attached evaluator still reads cannot be removed."
+  );
+  lines.push("\n## Evaluator Attachments\n");
+  lines.push(
+    "A saved evaluator attaches to a test suite (and, for checks that read only the conversation or the trace, to a run plan). After every scenario run the platform runs it and stores one result per evaluator on the run. A **required** pass/fail evaluator that fails fails the scenario; a score-only evaluator reports beside the verdict and never gates."
+  );
+  lines.push("```json");
+  lines.push('{ "evaluators": [{');
+  lines.push('  "evaluatorId": "<id from platform_list_evaluators>",');
+  lines.push('  "required": true,');
+  lines.push('  "mappings": {');
+  lines.push('    "output": { "type": "source", "sourceId": "trace", "path": ["tool_calls", "run_sql", "input"] },');
+  lines.push('    "expected_output": { "type": "source", "sourceId": "scenario", "path": ["fields", "golden_sql"] }');
+  lines.push("  }");
+  lines.push("}] }");
+  lines.push("```");
+  lines.push(
+    "Each mapping names where one evaluator input reads its value. Read the evaluator's inputs from `platform_get_evaluator`; every required input needs a mapping before a run. The paths:"
+  );
+  lines.push(
+    '- **conversation**: `["first_user_message"]`, `["last_agent_message"]`, `["transcript"]` (role: content lines) or `["messages"]` (JSON)'
+  );
+  lines.push(
+    '- **scenario**: `["situation"]`, `["criteria"]` or `["fields", "<identifier>"]` (a field the suite declares; not on a run plan)'
+  );
+  lines.push(
+    '- **trace**: `["contexts"]` (the retrieved contexts of the run\'s traces) or `["tool_calls", "<toolName>", "input" | "output"]` (the last matching tool call)'
+  );
+  lines.push(
+    '- a literal: `{ "type": "value", "value": "..." }`'
+  );
+  lines.push(
+    "A scenario whose field is blank skips the evaluators that read it, with a reason. A trace with no such tool call or no contexts fails them, with a reason. The results come back under `results.evaluations` on `platform_get_simulation_run`."
+  );
   lines.push("\n## Writing a Good Situation\n");
   lines.push(
     "The situation drives the user simulator. Include these elements:"
@@ -64,4 +116,4 @@ function formatScenarioSchema() {
 export {
   formatScenarioSchema
 };
-//# sourceMappingURL=discover-scenario-schema-UPBTCWPG.js.map
+//# sourceMappingURL=discover-scenario-schema-W7KBXA4J.js.map

package/dist/{discover-schema-ICGXPHUW.js → discover-schema-L6QDPON3.js}

@@ -1,4 +1,4 @@
-import "./chunk-BJ7B4R26.js";
+import "./chunk-SAS5OB7F.js";
 
 // src/schemas/filter-fields.ts
 var filterFields = [
@@ -445,4 +445,4 @@ function formatGroups() {
 export {
   formatSchema
 };
-//# sourceMappingURL=discover-schema-ICGXPHUW.js.map
+//# sourceMappingURL=discover-schema-L6QDPON3.js.map

package/dist/get-agent-7THF7HCU.js

@@ -0,0 +1,66 @@
+import {
+  getAgent
+} from "./chunk-EYXMLE5X.js";
+import "./chunk-GELJ2MQR.js";
+import "./chunk-WT4PO4A3.js";
+import "./chunk-SAS5OB7F.js";
+
+// src/tools/get-agent.ts
+var describeParameter = ({
+  name,
+  type,
+  options,
+  default: defaultValue,
+  description,
+  required
+}) => {
+  const parts = [type];
+  if (options?.length) parts.push(`one of ${options.join(", ")}`);
+  if (defaultValue !== void 0) parts.push(`default ${JSON.stringify(defaultValue)}`);
+  if (required) parts.push("required");
+  const suffix = description ? `: ${description}` : "";
+  return `- **${name}** (${parts.join(", ")})${suffix}`;
+};
+async function handleGetAgent({ id }) {
+  const agent = await getAgent(id);
+  const lines = [];
+  lines.push(`# ${agent.name}
+`);
+  lines.push(`**ID**: ${agent.id}`);
+  lines.push(`**Type**: ${agent.type}`);
+  if (agent.environment) lines.push(`**Environment**: ${agent.environment}`);
+  if (agent.status) lines.push(`**Status**: ${agent.status}`);
+  if (agent.owner?.name) lines.push(`**Owner**: ${agent.owner.name}`);
+  else if (agent.hostLabel) lines.push(`**Host**: ${agent.hostLabel}`);
+  if (agent.selectable === false) {
+    lines.push("**Run**: only its owner can run this agent");
+  }
+  if (agent.lastSeenAt) lines.push(`**Last seen**: ${agent.lastSeenAt}`);
+  lines.push(`**Created**: ${agent.createdAt}`);
+  lines.push(`**Updated**: ${agent.updatedAt}`);
+  if (agent.parameters && agent.parameters.length > 0) {
+    lines.push("\n## Parameters\n");
+    for (const parameter of agent.parameters) lines.push(describeParameter(parameter));
+  }
+  if (agent.instances && agent.instances.length > 0) {
+    lines.push(`
+## Instances (${agent.instances.length})
+`);
+    for (const instance of agent.instances) {
+      const label = instance.label ? ` (${instance.label})` : "";
+      lines.push(`- ${instance.hostname || instance.id}${label}, connected ${instance.connectedAt}`);
+    }
+  }
+  if (agent.config && Object.keys(agent.config).length > 0) {
+    lines.push("\n## Config\n");
+    lines.push("```json");
+    lines.push(JSON.stringify(agent.config, null, 2));
+    lines.push("```");
+  }
+  return lines.join("\n");
+}
+export {
+  describeParameter,
+  handleGetAgent
+};
+//# sourceMappingURL=get-agent-7THF7HCU.js.map

package/dist/{get-analytics-PVERWOA5.js → get-analytics-R6AZD3AW.js}

@@ -3,9 +3,9 @@ import {
 } from "./chunk-ZXKLPC2E.js";
 import {
   getAnalyticsTimeseries
-} from "./chunk-5UWY24RP.js";
-import "./chunk-KOO5733S.js";
-import "./chunk-BJ7B4R26.js";
+} from "./chunk-GELJ2MQR.js";
+import "./chunk-WT4PO4A3.js";
+import "./chunk-SAS5OB7F.js";
 
 // src/tools/get-analytics.ts
 function getGroupedData(bucket, groupBy) {
@@ -70,4 +70,4 @@ async function handleGetAnalytics(params) {
 export {
   handleGetAnalytics
 };
-//# sourceMappingURL=get-analytics-PVERWOA5.js.map
+//# sourceMappingURL=get-analytics-R6AZD3AW.js.map

package/dist/{get-dashboard-ZZPC3WMD.js → get-dashboard-P4WWVHFU.js}

@@ -1,9 +1,9 @@
 import {
   getDashboard
-} from "./chunk-UXXA4D6K.js";
-import "./chunk-5UWY24RP.js";
-import "./chunk-KOO5733S.js";
-import "./chunk-BJ7B4R26.js";
+} from "./chunk-S4GVX36W.js";
+import "./chunk-GELJ2MQR.js";
+import "./chunk-WT4PO4A3.js";
+import "./chunk-SAS5OB7F.js";
 
 // src/tools/get-dashboard.ts
 async function handleGetDashboard(params) {
@@ -27,4 +27,4 @@ async function handleGetDashboard(params) {
 export {
   handleGetDashboard
 };
-//# sourceMappingURL=get-dashboard-ZZPC3WMD.js.map
+//# sourceMappingURL=get-dashboard-P4WWVHFU.js.map

package/dist/{get-dataset-25TB2LQH.js → get-dataset-AKTYSQDI.js}

@@ -3,10 +3,10 @@ import {
 } from "./chunk-NUFPTHKU.js";
 import {
   getDataset
-} from "./chunk-J45SVWPR.js";
-import "./chunk-5UWY24RP.js";
-import "./chunk-KOO5733S.js";
-import "./chunk-BJ7B4R26.js";
+} from "./chunk-AN4OUA4D.js";
+import "./chunk-GELJ2MQR.js";
+import "./chunk-WT4PO4A3.js";
+import "./chunk-SAS5OB7F.js";
 
 // src/tools/get-dataset.ts
 function formatDatasetResponse(dataset) {
@@ -46,4 +46,4 @@ export {
   formatDatasetResponse,
   handleGetDataset
 };
-//# sourceMappingURL=get-dataset-25TB2LQH.js.map
+//# sourceMappingURL=get-dataset-AKTYSQDI.js.map