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

Annotations already declare readOnly=false, idempotent=true, destructive=false, so the safety profile is partly covered; the description adds what the annotations cannot: totals are recalculated after the edit, null clears a field, and certain field edits spawn a new conversion rather than mutating in place. The prompt-injection warning that invoice content is third-party UNTRUSTED data is substantive behavioral context for a write tool.

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 core action, the path syntax and the follow-up call are front-loaded before the caveats, and each sentence carries a distinct operational fact. The final security sentence is long, but it encodes a genuine risk instruction rather than padding.

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 3-parameter mutation tool with a nested free-form object and a known side effect, the description covers action, preconditions, follow-up, recomputation and injection risk; an output schema exists, so return values need not be described. Only authorization/permission requirements go unmentioned.

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 only 33%, and the description compensates for the weakest parameter: "changes" is a free-form path→value map in the schema, and the description supplies the path syntax with concrete examples ("buyer.siren", "invoice.issueDate", "lines.0.quantity") plus the null-to-clear semantics. It does not explain version/optimistic locking in prose, but that parameter is already self-documented in the schema.

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 uses a specific verb and resource: it corrects fields ("Corrige des champs") of a conversion in a defined state ("au statut needs_input"). It names the sibling it hands off to ("convert_to_facturx") and the exact continuation payload, so an agent can distinguish it from batch_convert or validate_facturx without opening any schema.

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?

It states the precondition (status needs_input or à régénérer), the required follow-up call with its exact argument shape, and an explicit do-not ("Ne décompte rien : relancez ensuite convert_to_facturx…"). It also gives the side-effect condition: editing a facture number, total TTC or client SIREN on an already-unlocked invoice creates a new conversion — a routing fact an agent could not infer.

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