Recherche dans le référentiel — Liste les clés publiques du référentiel, filtrables par domaine (fiscal/retraite) et/ou par mot-clé. Renvoie un résumé par clé (valeur, libellé, fiabilité) ; utiliser referentiel_valeur pour le détail sourcé complet. (sources: references_officielles.json v0.8.6)
| Name | Required | Description | Default |
|---|---|---|---|
| q | No | Mot-clé cherché dans la clé publique ou le libellé. | |
| domaine | No | Filtre par domaine. |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations are provided, so the description carries the full burden; it discloses the return shape (summary per key: valeur, libellé, fiabilité) and data provenance (references_officielles.json v0.8.6). It implicitly signals a read-only listing operation, but stops short of edge behaviors like pagination or ordering; still, this is substantive disclosure for a search 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?
One dense sentence front-loaded with the action, followed by a compact parenthetical source note. Every clause earns its place: scope, filters, return shape, sibling routing, and data provenance, with zero waste.
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 2-optional-param tool with no output schema and no annotations, the description covers the key gaps an agent needs: filter semantics, return shape, and the alternative for deeper detail. It omits the referentiel_versions sibling and the succession enum value from its domain list, which are minor gaps.
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 schema already fully documents q and domaine; the description adds no new parameter meaning. Minor inconsistency: it lists only fiscal/retraite as domains while the schema enum includes succession, though the schema remains authoritative.
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?
States a specific verb+resource ('Recherche dans le référentiel') and a precise scope: lists public keys filterable by domain and/or keyword. It explicitly differentiates itself from the sibling referentiel_valeur ('utiliser referentiel_valeur pour le détail sourcé complet'), so an agent can tell them apart without opening schemas.
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?
Provides an explicit routing rule: use referentiel_valeur when complete sourced detail is needed, which implies this tool is for quick key summaries. The filter conditions (domaine and/or q) are stated clearly, though it does not address the referentiel_versions sibling or give when-not conditions beyond the one alternative.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Add one secure layer between your agents and this server.
The tool families are clearly separated by domain prefixes (fiscal, retraite, referentiel), and the descriptions proactively distinguish near-neighbors such as fiscal_tmi vs fiscal_prelevements_sociaux and retraite_estimation vs retraite_pension_totale. Some initial confusion is possible between retraite_pension_regime and retraite_regimes, but the descriptions are detailed enough to resolve it.
Most tools follow a predictable snake_case domain_object pattern: fiscal_*, retraite_*, referentiel_*. The clear outlier is qotien_capacites, which uses a misspelled, non-domain prefix and breaks the otherwise consistent naming scheme.
With 18 tools, the server is heavier than the typical 3-15 sweet spot, though it is organized into recognizable fiscal, retirement, and referential clusters. The count is defensible for a two-domain server, but it begins to feel like a large MCP surface that agents must navigate carefully.
The server covers the main French income-tax and retirement calculation needs: income tax, marginal rate, social levies, CEHR, property capital-gains surtax, pension estimation, multi-regime totals, net pension, buybacks, progressive retirement, and parental surcharge. Minor gaps exist—such as no explicit long-career early-retirement tool and pensions_fournies being referenced rather than exposed as a first-class tool—but agents can work around them.