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

Annotations already declare this as a non-read-only, non-idempotent, non-destructive, closed-world operation, and the description adds real context beyond them: per-query cost (0.01 USDC), a hard cap of 20 queries per session, and that the session must be funded first. It omits what happens on a failed query (whether prepaid funds are consumed or refunded) and says nothing about the delivery_receipt or sandbox_receipt behavior, which matters for a spend-incurring call.

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?

Three compact sentences, each carrying distinct information: pricing/listing identity, the free alternative, and the required call sequence. The opening clause is slightly promotional ('first taste free') but still encodes the data_preview pointer, and the sequence is placed last where it functions as a checklist rather than preamble.

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 spend-incurring tool with no output schema and five parameters at 40% coverage, the description covers cost, quota, and sequencing but leaves gaps on return shape, error/refund behavior, and the meaning of the result-count and receipt parameters. An agent could invoke it correctly on the happy path but would be guessing at edge cases.

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 only 40%, so the description needs to compensate for the undocumented parameters, and it does not. Neither 'k' (1-10 result count), 'delivery_receipt', nor the 'query' string semantics are explained; session_id and sandbox_receipt rely entirely on the schema's own text. The only parameter-adjacent fact is the 20-queries-per-session cap.

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?

The description states a specific action and resource: per-query paid access to the 'msrwatch: Microsoft Research Blog' data listing at 0.01 USDC per query. It is distinguishable from siblings through the explicit 'first taste free via data_preview' contrast and the open → fund → query sequence. It loses a point only because the framing ('Buy per-query access') foregrounds the commercial model over the mechanical act of running a query.

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?

It names a concrete alternative for the free path (data_preview) and spells out the prerequisite sequence data_session_open → data_session_fund → data_session_query, which tells an agent what must exist before calling. It does not cover failure conditions, such as what to do with an unfunded or expired session, or when the escrow/funding-package siblings would be preferable.

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