Evaluate with Jev

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

Goes well beyond readOnlyHint/openWorldHint: questions run in parallel against the same state and cannot see each other's answers, provider is restricted to "typesafe", and images must live under JEV_IMAGE_DIRS or be inline data URLs with a 4 MiB cap and image-capable-model requirement. Those are exactly the behavioral facts an agent would otherwise discover by failing.

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

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

Front-loads the core purpose, then uses a bulleted question-type block, then tacks on optional concerns (images, provider). Dense and mostly waste-free, though the question-type bullets partially duplicate what the output schema already declares.

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

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

For a 5-parameter tool with deeply nested question objects, the description covers the state/questions interaction, per-type criteria contracts, model/image constraints, and even summarizes return fields despite an output schema existing. An agent has everything needed to construct a valid call on the first try.

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

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

Schema description coverage is 0%, so the description carries the full load and does so thoroughly: `state` as string or JSON, `questions` as an ID-keyed map with the three variants and their criteria shapes and limits (255 options, 2-10 levels), `images` formats and sources, `model` default, and `provider` restriction. Little about the parameters is left ambiguous.

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

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

States a specific verb and resource: ask typed questions about a piece of state, get calibrated probabilities back. That is far more precise than a name restatement. It does not, however, differentiate itself from siblings jev.noul, jev.choice, jev.score or jev.evaluate_batch, so an agent must infer that this is the multi-question, multi-type entry point.

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

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

Offers concrete operating guidance: one narrow judgment per question, question IDs are invisible to the model so `instructions` must be self-contained, backticked paths for structured state, never place images inside `state`, and images require an image-capable model. It stops short of saying when to prefer this over jev.noul / jev.choice / jev.score / jev.evaluate_batch.

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