Skip to main content
Glama

polaris_measurement_submissions

Submit measurements to be evaluated against the active fitness-function version. Supports single or batch submissions up to 100 items, with per-item status reporting.

Instructions

Push measurements for evaluation against the ACTIVE fitness-function version.

These are the ingest endpoints: they authenticate with the X-API-Key configured as POLARIS_INGEST_API_KEY (no Google login involved).

Actions and required parameters:

  • submit: fitness_function_id + submission — evaluates synchronously and returns the recorded evaluation (with replayed: true when a duplicate producerId+externalRunId delivery returned the original evaluation). Submissions against a non-active version yield 409.

  • submit_batch: items — up to 100 independent submissions, each {fitnessFunctionId, submission}; the 207 response carries per-item status (201 success, or 404/409/422/500 with an error message); per-item failures do not abort the batch.

submission is a JSON object with camelCase keys: required: fitnessFunctionVersion (the active version number); producerId (the producer declared by the active definition); externalRunId (producer-scoped run id, the idempotency natural key); observedAt (RFC 3339 UTC, not in the future, within the definition's maximum observation age); measurements (array of {criterionKey, value, unit, observedAt?}, one per criterion, duplicates rejected); optional: evaluationRequestId (request this submission fulfills), evidence (array of arbitrary evidence documents).

Input Schema

TableJSON Schema
NameRequiredDescriptionDefault
itemsNo
actionYes
submissionNo
fitness_function_idNo

Schema Changelog

Changes observed during successful MCP inspections.

  1. First observedv0.1.0

TDQS

A4.7/5.0
Behavior5/5

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

With no annotations present, the description carries the full burden and does so well: auth model (X-API-Key, no Google login), synchronous evaluation, idempotency via producerId+externalRunId, the `replayed: true` duplicate signal, 409 on non-active version, and batch 207 per-item semantics where failures don't abort. This is rich beyond what any structured field provides.

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-loads the purpose and auth model, then organizes actions and the submission payload into scannable bullets. It is long but nearly every line is load-bearing for a two-action ingest tool with an opaque payload; minor trimming is possible.

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?

No output schema and no annotations, yet the description covers return behavior (recorded evaluation, replayed flag, 207 per-item statuses) and error conditions. An agent has everything needed to construct a valid submission and interpret the response.

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

Parameters5/5

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

Schema coverage is 0% and all four parameters are opaque objects/strings, so the description must compensate fully. It documents which params each action needs and details the `submission` object's required and optional camelCase keys (including nested measurements structure and constraints like RFC 3339 UTC and maximum observation age), which the schema alone does not convey.

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 ('Push measurements for evaluation') and pins the scope to the ACTIVE fitness-function version. It also names the two ingest actions (submit, submit_batch), making it clearly distinguishable from sibling evaluation/measurement tools.

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?

Gives per-action required parameters and the conditions that select each one (submit vs submit_batch), plus the 409 behavior against non-active versions. It doesn't explicitly name sibling tools as alternatives, but the when-to-use guidance for each action is clear.

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