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

Annotations only cover the safety profile (readOnly, non-destructive, closed-world), and the description adds substantial behavior on top: cell scoring mirrors the Visibility page, runs are returned oldest first, the 40-cell ceiling, which inputs are rejected (unknown slice key, foreign topic_id, >4 runs), and the meaning of scopeTooSmall, noAnswers, notes/noteText, competitorsOmitted, and competitorSlice. That is exactly the extra context the annotations do not supply.

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?

Nearly every clause carries real information, and the hard limits come first. But it is delivered as one dense run-on paragraph mixing input rules, output semantics, competitor behavior, and error flags with no separation, which makes it slow to parse and pushes key operational facts into the middle.

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 5-parameter read tool with no output schema, the description covers everything an agent needs: valid input shapes, defaults, ordering, result layout, numeric ranges for AuraScore/Mention Rate/Share of Voice, and the degenerate cases (scopeTooSmall, noAnswers, scoring-rule notes). Nothing essential is left to guess.

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%, so 3 is the baseline, but the description goes well beyond it: it explains that a run is a run id or a YYYY-MM-DD day resolved to that day's latest completed run in time_zone, that prompt_type values map to Branded/Non-branded/Competitor-branded with a non_branded default, that filters need keys and value slugs from list-prompt-filters, and that include_competitors covers the first slice, top 30 by the newest run's Rank.

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 definition gives a specific verb plus resource ('Compare 2-4 visibility runs across 1-20 slices in one table') and pins down the output shape (rows = slices, column groups = runs). It never mentions the nearby siblings (get-visibility-score, get-visibility-summary, list-visibility-runs), so the agent must infer on its own that this is the multi-run comparison tool rather than the single-score tool.

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?

Hard boundaries are stated clearly (2-4 runs, 1-20 slices, 40-cell cap, >4 runs rejected), which implicitly tells the agent when the call is valid. But there is no explicit when-to-use/when-not guidance and no routing to alternatives such as list-visibility-runs for discovering run ids or get-visibility-summary for a single run, so usage is implied rather than directed.

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