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

Annotations already declare readOnlyHint=true and idempotentHint=true, so the safety profile is covered. The description adds substantial behavioral context beyond that: the 2026 barème with 2024 RFR thresholds, the tranche-selection logic tied to RFR and parts, the conditional 1% maladie on the complementary portion, and the output set. It also cites legal sources (CSS L136-8, L131-2, circulaire CNAV) for verifiability. This is rich, useful behavioral disclosure.

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 purpose ('Prélèvements sociaux sur une pension (brut → net)') is front-loaded, followed by the calculation detail, outputs, version, exclusion, and sources. It is dense but every clause carries information — the legal citations and barème version earn their place for a tax tool with a dated rate table. Slightly long, but justified by the tool's complexity.

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 complex, multi-rate French tax computation with no output schema, the description compensates well by stating the three return values (pension nette, tranche CSG, taux effectif), the applicable barème year, and the RFR reference year. The schema supplies parameter-level fallback behavior (RFR proxy → 'estimée'). Minor gaps remain — no edge-case behavior (e.g., zero pension) — but nothing an agent needs to invoke it correctly is missing.

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 rfr (including the proxy/estimation behavior when absent), parts (linear interpolation), montant_complementaire (affected regimes), and pension_annuelle_brute. The description adds genuine value on top by linking the CSG rate tiers to the RFR/parts parameters and explicitly tying the 1% maladie surcharge to montant_complementaire, which the schema's wording implies but does not state as a rate consequence.

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 a specific computation — prélèvements sociaux on a pension (brut → net) — and enumerates the exact components (CSG at 0/3.8/6.6/8.3%, CRDS, CASA, 1% maladie on complementary pension). It states the outputs (pension nette, tranche CSG, taux effectif) and explicitly scopes itself to 'Pensions de retraite uniquement', distinguishing it from fiscal_prelevements_sociaux. An agent can identify this tool without opening the 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?

The description gives an explicit when-to-use (retirement pension income) and when-not-to-use ('pour un revenu du capital → fiscal_prelevements_sociaux'), naming the alternative tool directly. This routing guidance is unambiguous and leaves nothing to inference.

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