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

Annotations already cover the read-only, idempotent, non-destructive profile, so the description's burden is lighter. It adds genuinely non-obvious behavior: revenue is attributed back to the day of the CLICK rather than the conversion day, and cells for cohort days not yet reached are null — both of which change how results should be read. No mention of cost/row limits or latency for a 62-parameter query.

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-loaded with the one-line purpose before the per-subject detail, and roughly 180 words is defensible for a tool with 62 optional parameters and no output schema. The sentences are long and heavily parenthetical, which costs some scannability, but little is filler.

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?

Covers the conceptual model, subject enumeration, split_by behavior, null semantics and output field meanings — the parts an agent cannot recover from the schema, and it needs to since no output schema exists. Gaps are secondary: no guidance on combining filters, on horizon defaults versus granularity, or on result size for a 62-parameter query.

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 89%, so the baseline is 3, and the description earns above it by adding meaning to the enums: what each subject actually cohorts on, what split_by produces (a second table whose rows are a dimension value, each with its own ladder/active_rate/returns_per_entity), and the horizon concept of 'payoff landed by day N'. Since there is no output schema, naming summary.arpu, returns_per_entity, active_rate, summary.ladder and life_days is a real contribution to interpreting results.

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?

Opens with a specific domain statement ('Cohort & LTV analytics') and then defines what a cohort is for each subject value (clicks, registration, ftd, push, pwa), which lets an agent tell it apart from generic reporting siblings like get_statistics or get_report. It never explicitly names those siblings as alternatives, so differentiation is implicit rather than stated.

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?

Gives strong subject-selection guidance: clicks for traffic-day ROI, registration for Lead-to-deposit, ftd for Sale-to-redeposit, push for subscriber base, pwa for installs. Also explains when split_by is worth setting (channel comparison against a day-row table). It stops short of stating when NOT to use this tool versus get_report/get_push_analytics.

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