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

Annotations already cover the safety profile (readOnlyHint=true, destructiveHint=false, openWorldHint=false), and 'Lecture seule' merely repeats them. The real added value is 'Indicatif — test d'infiltrometrie signe prime', which tells the agent results are advisory and a signed blower-door test overrides them — a behavioral caveat found nowhere in the structured fields. It does not, however, say what the check returns.

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 call condition, then packed with criteria codes and the scene shape in a single dense block, finished with the caveat and read-only note. Every clause carries information, though the JA-* codes and French domain jargon make it heavy to parse in one pass.

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 nested-object verification tool with no output schema, the input side is fully covered: criteria, allowed values, and the scene payload shape. What is missing is any indication of the result — whether it returns per-junction verdicts, a list of non-conformities, or a coverage summary — which the agent needs to interpret the call.

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?

The single parameter is at 100% schema coverage, so baseline would be 3. The description goes further by enumerating the accepted values the schema does not express as enums (type: mur_dalle/menuiserie/gaine/traversee; materiau: membrane/adhésif/calfeutrement) and by listing the per-junction sub-fields {nom, type, traite, materiau}.

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 and resource: 'VERIFIER L'ETANCHEITE A L'AIR PAR JONCTION (enveloppe)', and then enumerates exactly which criteria are checked (JA-TYPE, JA-TRAITEMENT, JA-MATERIAU, JA-COUVERTURE). The 'par jonction' qualifier implicitly separates it from a general airtightness check, but the sibling cao_bim_etancheite_air is never named, so the differentiation is inferred rather than stated.

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 devez VERIFIER L'ETANCHEITE A L'AIR PAR JONCTION (enveloppe)' gives a clear triggering condition, and the scoped criteria list makes the intended use concrete. No exclusions or alternative tools are named, so the agent is not told when to use cao_bim_etancheite_air or cao_bim_etancheite instead.

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