Skip to main content
Glama
SynapCores MCP Server

SynapCores MCP Server

Official

Train an AutoML model

train_model
Destructive

Train an AutoML model without SQL to predict a target column from other table columns, register the selected model under a name, and return its score and algorithm.

Instructions

Train an AutoML model that predicts one column of a table from its other columns, without writing SQL: wraps CREATE EXPERIMENT, runs max_trials trials, keeps the best algorithm, registers it as model_name, and returns the best score (accuracy for classification, R² for regression) and the algorithm. Then call predict with that name, or describe_model for full metrics. Reusing an existing model_name replaces that model; more trials take longer and can score higher. Table, target and feature names must use only letters, digits and underscores.

Tool call template

Replace each placeholder with a value matching the input schema. Only required parameters are included; add optional parameters as needed.

{
  "id": 1,
  "jsonrpc": "2.0",
  "method": "tools/call",
  "params": {
    "arguments": {
      "collection": "<collection>",
      "target": "<target>"
    },
    "name": "train_model"
  }
}

Input Schema

NameRequiredDescriptionDefault
targetYesColumn the model should predict.
featuresNoColumns to learn from; omit or leave empty to use every column except `target`.
task_typeNobinary_classification (default), multiclass_classification, regression, or classification (treated as binary).binary_classification
collectionYesTable that holds the training rows.
max_trialsNoHow many AutoML trials to run (default 3); more trials take longer and can find a better model.
model_nameNoName to register the model under (default {collection}_{target}_model); an existing model with this name is replaced.

Schema Changelog

Changes observed during successful MCP inspections.

  1. Changed6 schema fields changedv1.0.1
    • changedInput schema / properties / collection / description
      Previous value: -"Table/collection name containing training data"New value: +"Table that holds the training rows."
    • changedInput schema / properties / features / description
      Previous value: -"Optional list of feature columns. Empty/omitted uses all columns except target."New value: +"Columns to learn from; omit or leave empty to use every column except `target`."
    • changedInput schema / properties / max_trials / description
      Previous value: -"Number of AutoML trials (default: 3, higher = better but slower)"New value: +"How many AutoML trials to run (default 3); more trials take longer and can find a better model."
    • changedInput schema / properties / model_name / description
      Previous value: -"Optional model name (default: auto-generated as `{collection}_{target}_model`)"New value: +"Name to register the model under (default {collection}_{target}_model); an existing model with this name is replaced."
    • changedInput schema / properties / target / description
      Previous value: -"Column to predict"New value: +"Column the model should predict."
    • changedInput schema / properties / task_type / description
      Previous value: -"Task type (default: binary_classification)"New value: +"binary_classification (default), multiclass_classification, regression, or classification (treated as binary)."
  2. First observedv1.0.0

TDQS

A4.6/5.0
Behavior5/5

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

Adds real context beyond the annotations: it explains exactly why the tool is destructive ('Reusing an existing model_name replaces that model'), notes the cost/latency tradeoff ('more trials take longer and can score higher'), and discloses the naming restriction on collection/target/feature names. It also states the return contents, which the annotations alone would not convey.

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?

Purpose is front-loaded in the first clause, with mechanics, next steps, and caveats following in a logical order. It is a dense block of clauses in a single paragraph with some repetition of the max_trials cost point already in the schema, but little genuine waste.

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

Completeness5/5

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

For a six-parameter mutation tool with no output schema, the description covers what it returns (best score, accuracy or R² depending on task, plus algorithm), the destructive side effect, the cost model, and the follow-up tools. Nothing an agent needs to call it correctly appears to be missing.

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

Parameters4/5

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

Schema coverage is 100%, so the baseline is 3, but the description adds meaning the schema lacks: the identifier character restriction (letters, digits, underscores) and the framing of target/features as 'the column to predict' vs 'the other columns'. The max_trials and model_name semantics are largely restated from the schema, so it is not a full 5.

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 ('Train an AutoML model that predicts one column of a table from its other columns') and then details the mechanics: wraps CREATE EXPERIMENT, runs max_trials trials, keeps the best algorithm, registers it as model_name. That is clearly separable from describe_model, predict, and list_models.

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

Usage Guidelines4/5

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

Explicitly routes the agent forward: call `predict` with the registered name, or `describe_model` for full metrics. It does not, however, state when not to train (e.g., use query/execute for simple ad-hoc work) or any prerequisite such as existing table access.

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