A/B test sample size

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

Annotations already declare readOnly, idempotent, non-destructive and closed-world, so safety is covered. The description adds genuinely new behavioral information: that the tool works in two directions (solve for n given mde, or solve for mde given n) and that it is a pre-experiment planning aid rather than an analysis of observed data. It does not state the statistical assumptions baked in (α, power defaults) beyond what the schema shows.

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?

Dense and front-loaded — the two solve-for modes come first, then the required baseline inputs, then the usage rule. It reads as telegraphic notes ("Conversion: baseline_rate. Continuous: baseline_sd") rather than prose, which costs a little clarity on first read but wastes no sentences.

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 13-parameter planning tool with no output schema, the description omits a key decision rule: only metric is required, so which of mde / n_per_variant / daily_traffic / baseline_* must be supplied is never stated, and the shape of the result (per-variant n, total, days) is only implied. Coverage is adequate to call it but leaves the agent to infer valid parameter combinations from the schema alone.

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 coverage is only 31% across 13 parameters, but the description compensates well for the important ones: it explains mde semantics, that n_per_variant and mde are the two alternative solve-for inputs, that daily_traffic drives the days estimate, and that baseline_mean is only needed for a relative mde with continuous metrics. It adds a cross-parameter dependency (relative mde + continuous ⇒ baseline_mean) that the schema does not encode. Unexplained: alpha, variants, mde_type, alternative and adjust_for_variants are left to the schema.

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 ("Plan an A/B test") and immediately enumerates the two concrete outputs: visitors needed per variant for a given mde, and the smallest detectable lift for a given n_per_variant. It also names the metric branches (conversion vs continuous) and the derived days-needed output, so an agent knows exactly what it gets back without opening the schema.

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?

"Use in an exploration (planning a future test); a running hypothesis test sizes itself in step 4" gives an explicit when-to-use and a when-not-to-use. The exclusion is weaker than it could be because it refers to the alternative by workflow position ("step 4") rather than naming the sibling it competes with — most plausibly run_test and power_analysis — so the agent must infer the routing.

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