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

With readOnlyHint and openWorldHint already declaring this as a safe, closed-world read, the description still earns credit for a genuinely non-obvious semantic: "Counts everyone who matches, including addresses that are blocked or unsubscribed and will be skipped." That tells the agent the number is a raw match count, not a deliverable audience size. It passes there, but the justification for this being a 3 rather than higher is that the pagination mechanic and cumulative counting are covered more by the schema than by fresh behavioral disclosure.

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: the count-and-sample purpose arrives in the first clause before the usage advice and the paging caveat. Every sentence carries distinct information with no restating of the title. The phrase "with a sample of who they are" is a little loose and could specify sample size or fields.

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?

There is no output schema, so the description must carry return-value semantics, and it does cover the essentials: a count, a sample, is_complete, and a cursor. The counting-inclusion caveat is exactly the kind of detail an agent would otherwise get wrong when reporting to a user. Minor gaps remain around what the sample actually contains and whether the count is segmented in any way.

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 67%, and the filter_groups parameter already carries a precise matching-semantics description in the schema itself (a contact passes every filter of at least one group; an empty list means everyone). The description adds the one thing the schema does not: "the counts of the calls add up," clarifying that paged results are cumulative rather than per-page. project_id remains entirely undocumented in both places, which keeps this at the baseline rather than above it.

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 opens with a specific verb and resource: "Count the contacts a set of filter groups matches right now, with a sample of who they are." That is far more precise than the title alone and clearly separates it from mutation tools like create_campaign. It never names or contrasts with the nearest sibling, search_contacts, so differentiation from that read-side tool is left to inference.

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?

"Show it to the user before a campaign goes live or a broadcast sends" gives a concrete when-to-use condition tied to the user's workflow. The pagination rule ("When is_complete is false, call again with the returned cursor") is explicit. What's missing is a when-not: nothing says to avoid it for one-off contact lookups where search_contacts is a better fit.

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