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

With annotations already declaring readOnlyHint=true, the description need not repeat that it is read-only. It adds meaningful behavioral context: it identifies the specific data coverage ('dernier exercice exploitable', 'division NAF', 'tranche de chiffre d'affaires'), states that anonymous access excludes individual companies, discloses the cost (1 quota unit), and explains the output when no comparison is possible. It does not detail pagination or response format, but for a single-id query with no output schema that is a minor gap.

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?

The description is dense and information-packed while remaining under ~90 words. It opens with the main action, lists ratios, explains the output, and finishes with the access limitation and cost. Every sentence adds value; the only slight criticism is that it is a long single paragraph that could be more scannable with bullets or sentence breaks, but it is still well-structured for an API description.

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 read-only, single-parameter tool, the description is quite complete: it explains what it compares, how it compares, what the output contains, the access restriction, and the quota cost. The main missing detail is exact output schema/format, but since the tool has no output schema and lists the outputs textually, the description covers the essentials. It could also mention the expected data source or that results are based on the latest available fiscal data, but the description already says 'dernier exercice exploitable'.

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?

Even though schema coverage is 100%, the description adds real meaning to the only parameter: it explains that the SIREN is for a French company, that comparison is by sector/taille, and provides a concrete JSON example. The description thus helps the agent pass a valid SIREN and interpret the response, well beyond the schema's simple pattern and example.

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 states a specific verb ('Compare') and a precise resource (ratios of a French company's last exploitable fiscal year against sector peers by NAF division and, when possible, size bracket). It enumerates the exact ratios and outputs (value, quartiles, position, comparison level, number of comparable companies, or reason for failure). This clearly distinguishes it from the sibling tools (e.g., propecto_finances for raw financials, propecto_marche for market data).

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?

The description effectively explains when to use the tool: when you need a sectoral positioning of a French company's ratios. It also gives a clear exclusion condition: 'Sans clé API (entrée anonyme), les entreprises individuelles ne sont pas servies' – when the user has no API key/anonymous entry. It names sibling tools only implicitly by the function's position among propecto_* tools, but it doesn't explicitly say 'use propecto_finances instead for raw financials'. It is strong on entry requirements and failure modes.

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