Bernstein - Multi-agent orchestration

Overview Schema Source Related Servers Discussions Admin

bernstein_run

Queues a multi-agent orchestration run that writes project state and starts agent work from a stated goal, returning a task_id and run_id to poll for progress instead of waiting.

Instructions

Start an orchestration run. A run does real work and takes minutes to hours; the call returns once the run is queued, not when it finishes. Do not re-issue it while waiting, that starts a second run. Follow the run with bernstein_run_status, passing either the returned task_id or the returned run_id, after waiting the returned poll_after_ms. Pass parent_task_id to create the run as a subtask of an existing task. The queued orchestration writes project state and starts agent work. Host effects: writes files; spawns agent processes; makes network requests. Set template=administrative to create an empty audited manual container without a planner or agent processes; goal may be omitted for this template.

Input Schema

Name	Required	Default
`goal`	No
`role`	No
`scope`	No
`priority`	No
`template`	No	orchestration
`complexity`	No
`parent_task_id`	No
`estimated_minutes`	No

Output Schema

Name	Required	Description	Default
`_meter`	Yes
`result`	Yes

Schema Changelog

Changes observed during successful MCP inspections.

Changed5 schema fields changedv3.23.0Oct 26, 2026

addedInput schema / anyOf

Added value: +[
+  {
+    "required": [
+      "goal"
+    ]
+  },
+  {
+    "properties": {
+      "template": {
+        "const": "administrative"
+      }
+    },
+    "required": [
+      "template"
+    ]
+  }
+]

changedInput schema / description

Previous value: -"Start an orchestration run. A run does real work and takes minutes to hours; the call returns once the run is queued, not when it finishes. Do not re-issue it while waiting, that starts a second run. Follow the run with bernstein_run_status, passing either the returned task_id or the returned run_id, after waiting the returned poll_after_ms. Pass parent_task_id to create the run as a subtask of an existing task. The queued orchestration writes project state and starts agent work. Host effects: writes files; spawns agent processes; makes network requests."New value: +"Start an orchestration run. A run does real work and takes minutes to hours; the call returns once the run is queued, not when it finishes. Do not re-issue it while waiting, that starts a second run. Follow the run with bernstein_run_status, passing either the returned task_id or the returned run_id, after waiting the returned poll_after_ms. Pass parent_task_id to create the run as a subtask of an existing task. The queued orchestration writes project state and starts agent work. Host effects: writes files; spawns agent processes; makes network requests. Set template=administrative to create an empty audited manual container without a planner or agent processes; goal may be omitted for this template."

addedInput schema / properties / template

Added value: +{
+  "default": "orchestration",
+  "enum": [
+    "orchestration",
+    "administrative"
+  ],
+  "type": "string"
+}

removedInput schema / required
```
Removed value: -[
-  "goal"
-]
```

changedOutput schema / properties / result / anyOf

Previous value: -[
-  {
-    "additionalProperties": false,
-    "properties": {
-      "parent_task_id": {
-        "type": "string"
-      },
-      "poll_after_ms": {
-        "type": "integer"
-      },
-      "run_id": {
-        "type": "string"
-      },
-      "status": {
-        "type": "string"
-      },
-      "task_id": {
-        "type": "string"
-      },
-      "title": {
-        "type": "string"
-      }
-    },
-    "required": [
-      "task_id",
-      "title",
-      "status",
-      "run_id",
-      "poll_after_ms"
-    ],
-    "type": "object"
-  },
-  {
-    "additionalProperties": true,
-    "properties": {
-      "error": {
-        "type": "string"
-      },
-      "hint": {
-        "type": "string"
-      }
-    },
-    "required": [
-      "error"
-    ],
-    "type": "object"
-  }
-]New value: +[
+  {
+    "additionalProperties": false,
+    "properties": {
+      "parent_task_id": {
+        "type": "string"
+      },
+      "poll_after_ms": {
+        "type": "integer"
+      },
+      "run_id": {
+        "type": "string"
+      },
+      "status": {
+        "type": "string"
+      },
+      "task_id": {
+        "type": "string"
+      },
+      "title": {
+        "type": "string"
+      }
+    },
+    "required": [
+      "task_id",
+      "title",
+      "status",
+      "run_id",
+      "poll_after_ms"
+    ],
+    "type": "object"
+  },
+  {
+    "description": "Authenticated task aggregate returned by administrative run surfaces.",
+    "properties": {
+      "chainHead": {
+        "title": "Chainhead",
+        "type": "string"
+      },
+      "chainVerified": {
+        "const": true,
+        "title": "Chainverified",
+        "type": "boolean"
+      },
+      "goal": {
+        "title": "Goal",
+        "type": "string"
+      },
+      "receiptHash": {
+        "title": "Receipthash",
+        "type": "string"
+      },
+      "runId": {
+        "title": "Runid",
+        "type": "string"
+      },
+      "status": {
+        "enum": [
+          "working",
+          "completed"
+        ],
+        "title": "Status",
+        "type": "string"
+      },
+      "tasks": {
+        "items": {
+          "additionalProperties": {
+            "type": "string"
+          },
+          "type": "object"
+        },
+        "title": "Tasks",
+        "type": "array"
+      },
+      "template": {
+        "const": "administrative",
+        "default": "administrative",
+        "title": "Template",
+        "type": "string"
+      },
+      "tenant_id": {
+        "title": "Tenant Id",
+        "type": "string"
+      }
+    },
+    "required": [
+      "runId",
+      "goal",
+      "tenant_id",
+      "status",
+      "tasks",
+      "chainHead",
+      "chainVerified",
+      "receiptHash"
+    ],
+    "title": "AdministrativeRunStatus",
+    "type": "object"
+  },
+  {
+    "additionalProperties": true,
+    "properties": {
+      "error": {
+        "type": "string"
+      },
+      "hint": {
+        "type": "string"
+      }
+    },
+    "required": [
+      "error"
+    ],
+    "type": "object"
+  }
+]

Changed1 schema field changedv3.15.0Aug 14, 2026

changedInput schema / description

Previous value: -"Start an orchestration run. A run does real work and takes minutes to hours; the call returns once the run is queued, not when it finishes. Do not re-issue it while waiting, that starts a second run. Follow the run with bernstein_run_status, passing either the returned task_id or the returned run_id, after waiting the returned poll_after_ms. Pass parent_task_id to create the run as a subtask of an existing task."New value: +"Start an orchestration run. A run does real work and takes minutes to hours; the call returns once the run is queued, not when it finishes. Do not re-issue it while waiting, that starts a second run. Follow the run with bernstein_run_status, passing either the returned task_id or the returned run_id, after waiting the returned poll_after_ms. Pass parent_task_id to create the run as a subtask of an existing task. The queued orchestration writes project state and starts agent work. Host effects: writes files; spawns agent processes; makes network requests."

Changed33 schema fields changedv3.11.0Jul 28, 2026

addedInput schema / $schema

Added value: +"http://json-schema.org/draft-07/schema#"

addedInput schema / additionalProperties
```
Added value: +false
```

addedInput schema / description

Added value: +"Start an orchestration run. A run does real work and takes minutes to hours; the call returns once the run is queued, not when it finishes. Do not re-issue it while waiting, that starts a second run. Follow the run with bernstein_run_status, passing either the returned task_id or the returned run_id, after waiting the returned poll_after_ms. Pass parent_task_id to create the run as a subtask of an existing task."

removedInput schema / properties / complexity / default
```
Removed value: -"medium"
```

addedInput schema / properties / complexity / enum

Added value: +[
+  "low",
+  "medium",
+  "high"
+]

removedInput schema / properties / complexity / title
```
Removed value: -"Complexity"
```
removedInput schema / properties / estimated_minutes / default
```
Removed value: -30
```
addedInput schema / properties / estimated_minutes / maximum
```
Added value: +100000
```
addedInput schema / properties / estimated_minutes / minimum
```
Added value: +0
```
removedInput schema / properties / estimated_minutes / title
```
Removed value: -"Estimated Minutes"
```
addedInput schema / properties / goal / maxLength
```
Added value: +8192
```
addedInput schema / properties / goal / minLength
```
Added value: +1
```
removedInput schema / properties / goal / title
```
Removed value: -"Goal"
```

addedInput schema / properties / parent_task_id

Added value: +{
+  "maxLength": 256,
+  "minLength": 1,
+  "pattern": "^[A-Za-z0-9_.:-]+$",
+  "type": [
+    "string",
+    "null"
+  ]
+}

removedInput schema / properties / priority / default
```
Removed value: -2
```
addedInput schema / properties / priority / maximum
```
Added value: +3
```
addedInput schema / properties / priority / minimum
```
Added value: +1
```
removedInput schema / properties / priority / title
```
Removed value: -"Priority"
```
removedInput schema / properties / role / default
```
Removed value: -"backend"
```
addedInput schema / properties / role / maxLength
```
Added value: +64
```
addedInput schema / properties / role / minLength
```
Added value: +1
```
removedInput schema / properties / role / title
```
Removed value: -"Role"
```
removedInput schema / properties / scope / default
```
Removed value: -"medium"
```

addedInput schema / properties / scope / enum

Added value: +[
+  "small",
+  "medium",
+  "large"
+]

removedInput schema / properties / scope / title
```
Removed value: -"Scope"
```

changedInput schema / title

Previous value: -"bernstein_runArguments"New value: +"bernstein_run"

addedOutput schema / additionalProperties
```
Added value: +false
```

addedOutput schema / properties / _meter

Added value: +{
+  "additionalProperties": false,
+  "properties": {
+    "call_id": {
+      "type": "string"
+    },
+    "cost_usd": {
+      "type": "number"
+    },
+    "error": {
+      "type": "string"
+    },
+    "latency_ms": {
+      "type": "number"
+    },
+    "ok": {
+      "type": "boolean"
+    },
+    "tool": {
+      "type": "string"
+    },
+    "ts": {
+      "type": "string"
+    }
+  },
+  "required": [
+    "tool",
+    "call_id",
+    "latency_ms",
+    "cost_usd",
+    "ok",
+    "ts"
+  ],
+  "type": "object"
+}

addedOutput schema / properties / result / anyOf

Added value: +[
+  {
+    "additionalProperties": false,
+    "properties": {
+      "parent_task_id": {
+        "type": "string"
+      },
+      "poll_after_ms": {
+        "type": "integer"
+      },
+      "run_id": {
+        "type": "string"
+      },
+      "status": {
+        "type": "string"
+      },
+      "task_id": {
+        "type": "string"
+      },
+      "title": {
+        "type": "string"
+      }
+    },
+    "required": [
+      "task_id",
+      "title",
+      "status",
+      "run_id",
+      "poll_after_ms"
+    ],
+    "type": "object"
+  },
+  {
+    "additionalProperties": true,
+    "properties": {
+      "error": {
+        "type": "string"
+      },
+      "hint": {
+        "type": "string"
+      }
+    },
+    "required": [
+      "error"
+    ],
+    "type": "object"
+  }
+]

removedOutput schema / properties / result / title
```
Removed value: -"Result"
```
removedOutput schema / properties / result / type
```
Removed value: -"string"
```

changedOutput schema / required

Previous value: -[
-  "result"
-]New value: +[
+  "result",
+  "_meter"
+]

removedOutput schema / title
```
Removed value: -"bernstein_runOutput"
```

Addedv2.7.0May 25, 2026
Removedv2.6.0May 22, 2026
Addedv2.2.0May 18, 2026
Removedv2.1.0May 18, 2026
Addedv1.10.8May 13, 2026
Removedv1.10.7May 11, 2026
First observedv0.1.0Apr 21, 2026

TDQS

A4.4/5.0

Behavior5/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

No annotations exist, so the description carries the entire burden and does so: it discloses async completion semantics, non-idempotency, host effects (file writes, agent process spawning, network requests), and the differing behavior of the administrative template (no planner, no agent processes, empty audited container). These are exactly the traits an agent could not infer from the schema.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness4/5

Is the description appropriately sized, front-loaded, and free of redundancy?

Front-loaded with the core async contract, then sequential guidance, then effects, then the template variant — each sentence carries information. It is dense and slightly long, and it repeats the schema's own description verbatim, but nothing is padded.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness4/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

For a complex, side-effecting async tool with no annotations, the description covers timing, idempotency, follow-up flow, and host effects, and the output schema covers the returned task_id/run_id/poll_after_ms. The remaining gap is the undocumented tuning parameters, which leaves the definition short of fully complete.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters2/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema description coverage is 0%, so the description must explain all 8 parameters, but it only covers template, parent_task_id, and the goal-optionality constraint. role, scope, priority, complexity, and estimated_minutes get no guidance at all, leaving an agent to guess their effect on the run.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

States a specific verb and resource ('Start an orchestration run') and immediately defines what a run is: real work lasting minutes to hours, returning when queued. It is clearly separable from bernstein_run_status, bernstein_create_task, and the other siblings because it names the follow-up tool and contrasts run vs. status.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines5/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

Explicit when-and-how guidance: do not re-issue while waiting (a second run starts), follow with bernstein_run_status using the returned task_id or run_id after poll_after_ms, and use parent_task_id to nest as a subtask. It also gives a distinct usage branch for template=administrative where goal may be omitted — rare, actionable routing detail.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

Deploy Server