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

Annotations already cover read-only, non-destructive, closed-world, so the bar is lower, yet the description adds genuinely new behavior: scoring rules (0-100 per criterion, family average, Excellent >= 80 / Très bon >= 60 / Bon >= 40) and, importantly, that criteria without declared data stay 'non évalués' and do not inflate the score. The 'Indicatif — la certification réelle relève d'un assesseur accrédité' disclaimer is meaningful caveat context. Minor redundancy: 'Lecture seule' merely repeats readOnlyHint.

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?

Purpose is front-loaded and the family/criteria list is dense and informative rather than padded. The trailing 'Lecture seule' duplicates the annotation, and the disclaimer sentence, while valuable, could have been folded into the lead. Overall well-structured for its 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?

With no output schema, the description carries the return-shape burden and does so well: families, criteria, per-criterion score, family average and global level are all spelled out, plus the 'non évalués' rule. The remaining gap is the missing referentiel semantics and the lack of any statement about how the scene must be populated for criteria to be evaluated.

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 100%, so the baseline is 3; the schema already documents all three parameters including enum-ish values. The description clarifies what 'certificat' adds (recommandations), but never mentions the 'referentiel' parameter (hqe/breeam/leed) at all — an agent cannot learn from the description how choosing a referentiel changes the families, criteria or thresholds, which is the one place added semantics would have mattered.

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 concrete operation (consolidate/produce the environmental certification scorecard of a scene) and enumerates the exact output content: 4 families with their criteria, 0-100 per criterion, family averages, and global level thresholds. This is far beyond restating the title. However, it never names or contrasts itself with the many sibling criterion tools it presumably aggregates (cao_bim_etancheite, cao_bim_acv, cao_bim_qai, cao_bim_eclairage, etc.), so the boundary is left implicit.

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?

'Appelez quand vous voulez le scorecard...' establishes intent but is close to circular — it restates the output rather than a task context. The two modes are distinguished (action=evaluation default vs certificat 'avec recommandations'), which is real guidance, but there is no when-not, no prerequisite (e.g. the scene must already carry the per-lot business data), and no mention of which sibling tools feed or replace it.

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