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

Annotations already cover the safety profile (readOnly, non-destructive, closed-world), so credit goes to the extra context: 'it spends nothing', the warning that network-attributed numbers are 'never added to Marsell's count', and the partial-data state when 'counting.on false' where spend and network numbers survive but Marsell's own counts do not. That is meaningful beyond annotations, though it never says what the response shape or ordering looks like.

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 payload, which is right, but it is one long run-on sentence chaining results, network caveats, per-visit attribution and a degradation mode. The 'counting.on false:' fragment reads like leaked code and breaks the flow. Nearly every clause earns its place, but the packaging makes it harder to scan than a bulleted or paragraph-split version.

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?

No output schema exists, so the description must convey the return shape — it does, enumerating the metrics reported and the accompanying per-visit attribution section. It also discloses the important degraded mode (counting disabled) and the double-counting hazard. Only the pagination/ordering of the per-visit list and identity of the network are left implicit.

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 100% — both parameters already document the 1–365 range, the 90-day default and the project id-or-exact-name rule. The description only echoes the date window ('for the last days') and adds no format, unit or interaction detail. Baseline 3 applies since the schema carries the load.

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 (reports) and a well-bounded resource: per-ad and per-campaign visits, accounts opened, paid subscriptions, money net of refunds, spend and cost per result, plus the network's separate claims and per-visit attribution. It also distinguishes itself from the sibling that covers spend only ('Every ad on Sell with what it spends this month is place_open sell'). An agent can tell exactly what this returns without opening anything else.

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?

'Read it when the owner asks what an ad or campaign brought, or before saying an ad worked' gives a concrete trigger, and it routes a narrower spend-only need to place_open sell. What's missing is an explicit exclusion — e.g. when to use get_analytics or metrics_series instead — but the positive guidance is clear and actionable.

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