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

Annotations already cover the safety profile (readOnly, idempotent, non-destructive, closed-world), so the bar is lower, and the description adds a genuinely non-obvious behavioral rule: this is a dispatch facade whose per-operation contract must be fetched via decrire_outil first. That is exactly the kind of context annotations cannot express. It still omits whether operations are pure reads at the sub-level and how failures are surfaced.

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 domain sentence is front-loaded and compact, but the final two sentences say the same thing twice ('appeler decrire_outil ... avant usage' / 'Appeler decrire_outil avec nom = opération choisie ... avant usage'). Roughly a third of the text is redundant repetition rather than added information.

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 12-operation dispatcher with a free-form arguments bag, no output schema, and zero parameter documentation, the description does address the single biggest risk by mandating contract discovery via decrire_outil. It still leaves the operation catalogue unexplained and gives no sense of what a result looks like or how the 12 operations partition the domain.

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%: the `operation` enum lists 12 opaque French operation names with no descriptions, and `arguments` is a free-form object with additionalProperties and no documented shape. The description compensates partially by directing the agent to decrire_outil for the exact contract of the chosen operation, but it never explains what `arguments` should contain or that it varies per operation.

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 names the topical domain (French case law, rapporteur public conclusions, relations between decisions) but never states what the tool mechanically does — it is a router over 12 operations, and no verb or resource framing conveys that. It gestures at differentiation ('distinguer chaque juridiction et formation') without naming any sibling explorer it is distinct from. An agent knows the subject area but not the action.

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 two-step protocol is explicit and repeated: pick an operation, then call decrire_outil with that operation name to get its exact contract before use. That is real, actionable guidance. However, nothing says when to reach for this explorer versus explorer_procedure_penale, explorer_qualification_et_peines, or the other domain siblings, nor how to choose among the 12 listed operations.

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