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

With no annotations, the description carries the full burden and does well: it discloses that keyword matching may resolve to multiple options and surface a listing error, and that ImagePME can refuse output for statistical secrecy (small sample), with a suggested remedy. It does not mention authentication, rate limits, or whether downloads are cached/overwritten.

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 Args/Returns/Raises structure is front-loaded and scannable, and each section earns its place given the parameter count. It is somewhat long, but the length is justified by six undocumented parameters and non-trivial error behavior.

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 six-parameter tool with 0% schema coverage, no annotations, and a sibling, the description covers purpose, all parameter semantics, error conditions, and remediation. Even though an output schema exists (so Returns is not strictly necessary), the added field-level detail is harmless and helpful.

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 0% and there are 6 parameters, so the description must compensate fully — and it does, documenting every parameter: default values, the allowed niveau_sectoriel values, required-if conditions for secteur/territoire, loose-matching tolerance (NAF codes, keywords, full labels), and format options.

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 ('Télécharge') and resource ('indicateurs TDFC, données fiscales annuelles') from a named source (ImagePME). It is naturally distinguished from the sibling get_indicateurs_tva by the TDFC/TVA subject matter, though the description never names the sibling explicitly.

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 clear conditional context: secteur is required when niveau_sectoriel != 'tous', territoire when niveau_geo != 'national', and it advises retrying with a broader geographic/sectoral level on secret-statistique errors. It does not, however, route the agent between this tool and get_indicateurs_tva.

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

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

With no annotations, the description carries the full burden and does so richly: it discloses that each response returns ~12 months / ~11 trimesters of history, that July+August are merged as '07+08', the per-request latency, and that results can be refused under secret statistique with a recommended widening retry. Nothing about the operation's runtime behavior is hidden.

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 core definition, then the key usage warning, then Args/Returns/Raises — a clean structure where every block adds information. It is on the long side, and the Args section partly restates parameter names already in the schema, but the added semantic detail justifies the length.

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 an 8-parameter tool with zero schema coverage, no annotations and an output schema, the description supplies everything needed: conditional requirements, accepted value formats, the return shape (file path plus Excel table with échantillon/ICA/ICAC per period) and error handling. An agent can invoke it correctly without further inference.

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 description coverage is 0%, so the description must compensate, and it does: it enumerates valid values for periodicite, niveau_sectoriel, niveau_geo and format, gives an annee example, and explains that secteur/territoire accept NAF codes, keywords or full labels with fuzzy-matching plus a disambiguation error path. Conditional requirements (secteur required unless niveau_sectoriel='tous') are stated explicitly.

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 resource and metric: 'Indicateurs TVA (ICA/ICAC) d'ImagePME : variation du chiffre d'affaires', and defines both indicators (ICA = variation vs. same period last year, ICAC = cumulative year-to-date). It does not name or differentiate itself from the sibling get_indicateurs_tdfc, so it falls short of a 5.

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?

Explicitly tells the agent how to query economically ('une seule requête suffit', 'interroge UNIQUEMENT la période la plus récente'), warns against a request per period, and gives a ~1 minute cost. It also prescribes recovery behavior for both failure modes (ambiguous keyword, secret statistique), which is exactly when-to-use guidance.

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