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

Annotations already cover readOnly/idempotent/non-destructive, so the '[read-only]' tag is redundant, but the description adds real value: it discloses the response structure (per-arm n, leading variant) and a statistical caveat about small samples. The wrapped REST endpoint is a useful traceability detail.

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?

Two tight sentences, front-loaded with the read-only tag and the payload shape. The parenthetical endpoint note is marginal but small; nothing else is wasted.

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?

With no output schema and 0% schema coverage, the description does well on return shape but leaves the window parameter and the include flag unaddressed, which is a meaningful gap for a results tool where time window changes the answer.

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 must carry the parameter burden and largely does not. It never explains what 'window' means or why the default is 48h versus 7d/latest, and it only implicitly hints at 'include=posts' via the word 'posts'.

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?

Names a specific verb+resource and describes the shape of the payload precisely: posts split by compareKey, leading variant by primary metric, n per arm. That distinguishes it from the metadata-oriented get_experiment, though the sibling is not named explicitly.

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?

The 'Small samples: read medians with their n' line gives interpretive guidance but not selection guidance – it never says when to call this versus get_experiment or list_experiments, nor any preconditions (e.g. experiment must have concluded or have traffic).

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