Skip to main content
Glama

Score predictions against the sealed test labels

score

Scores a CSV or Parquet predictions file against a dataset’s sealed test set and returns aggregate metrics only, without exposing labels or per-row results.

Instructions

Score a predictions file (CSV or Parquet) on the sealed test set.

The file needs the dataset's id column and a 'prediction' column. Returns aggregate metrics only, never labels or per-row results. Spends one reveal when it succeeds.

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": {
      "dataset": "<dataset>",
      "predictions_path": "<predictions_path>"
    },
    "name": "score"
  }
}

Input Schema

NameRequiredDescriptionDefault
seedNo
datasetYes
predictions_pathYes

Output Schema

NameRequiredDescriptionDefault
modeYes
stampNoProvenance: commit, dirty flag, data/scorer/lockfile hashes, seed, ledger entry
budgetNo
statusYes'scored' means metrics were revealed and one reveal was spent
datasetYes
messageYesPlain-English summary of what happened
metricsNo
findingsNo
n_scoredNo
remainingNoReveals left after this call

Schema Changelog

Changes observed during successful MCP inspections.

  1. First observedv0.1.0

TDQS

A3.5/5.0
Behavior4/5

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

With no annotations, the description carries the full burden and does disclose two real behavioral traits: output is restricted to aggregate metrics with no labels or per-row results (leakage protection), and a successful call consumes a reveal budget. It omits what happens on failure (does a failed call still spend a reveal?) and whether re-scoring the same file costs again, which keeps it from a 5.

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?

Three short sentences, front-loaded with the action and followed by the input requirement and the two behavioral caveats. No filler; the only small weakness is that the format constraints sit in a separate paragraph rather than inline with the file description.

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?

An output schema exists, so return shape needn't be described, and the description still adds the important caveat that only aggregate metrics come back. For a budget-spending scoring call with no annotations, the main remaining gap is failure/retry semantics and the role of seed, but the essential contract is present.

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 compensate for all three parameters, and it only meaningfully does so for predictions_path (CSV or Parquet, must contain the dataset's id column and a 'prediction' column). The required 'dataset' parameter and the optional 'seed' parameter are never explained, including what seed actually controls in scoring.

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

Purpose4/5

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

States a specific verb and resource ('Score a predictions file') plus the target scope ('the sealed test set'), so an agent knows exactly what the operation does. It does not name a sibling (e.g. evaluate or check_leakage) to differentiate itself, so it stops short of a 5.

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

Usage Guidelines3/5

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

Usage is only implied: the agent must infer that this is the tool for turning an already-written predictions file into metrics, and that alternatives like evaluate are for something else. The cost note ('Spends one reveal when it succeeds') gives a reason to be deliberate, but there is no explicit when-to-use vs when-not framing.

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