QUAND une facture doit être conforme à la facturation électronique française (réforme e-facture, EN 16931). Facture Factur-X (PDF/A-3 + XML EN 16931). POST /api/v1/documents/facturx. seller{}, buyer{}, lines[] (description, quantity, unit_price HT, vat_rate). format=preview pour PDF sans XML.
| Name | Required | Description | Default |
|---|---|---|---|
| buyer | Yes | ||
| lines | Yes | ||
| notes | No | ||
| format | No | ||
| seller | Yes | ||
| due_date | No | ||
| invoice_date | Yes | ||
| invoice_number | Yes |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations are provided, so description carries full burden for behavioral disclosure. Does not mention side effects (e.g., deletion, modification), required permissions, rate limits, or idempotency. Describes creation of a document but not whether it is saved externally or returned inline. Lacks transparency expected for a mutation tool with zero annotation coverage.
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?
Description is concise (single sentence with abbreviations) but front-loads condition and type. Contains run-on structure with mixed information (condition, format, endpoint, params). Could be split into paragraphs for clarity, but remains information-dense without redundancy.
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 tool with 8 parameters (5 required, nested objects, no output schema), the description leaves gaps: no return value explanation, no example usage, no error conditions. Sibling list is long but no explicit contrast. Adequate for domain experts but insufficient for AI agent to infer all invocation details.
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 0%. Description partially compensates by listing fields for lines[] (description, quantity, unit_price HT, vat_rate) and explaining format enum (facturx vs preview). However, misses explanation for 3 required parameters (invoice_number, invoice_date) and optional ones (notes, due_date). Does not define nested objects (buyer, seller).
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?
Description starts with 'QUAND' specifying the exact use case (French electronic invoicing reform, EN 16931). It names the output format (PDF/A-3+XML) and distinguishes from sibling tools (e.g., devis_travaux, quittance_loyer) by standard compliance. The tool name and content align precisely.
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?
Explicitly states when to use: 'QUAND une facture doit être conforme à la facturation électronique française'. Provides conditional guidance via format=preview for alternative output. Does not explicitly state when not to use, but the condition implies exclusion of non-compliant invoices. Sibling context reinforces differentiation.
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.
Most tools are clearly distinct by purpose (e.g., quittance_loyer vs relance_impaye, or relance_facture_client vs relance_impaye). A few pairs like bon_intervention and fiche_mission could cause confusion, but detailed descriptions help resolve ambiguity.
Tool names mix French and English, and use inconsistent patterns: some are noun-based (annonce_location, quittance_loyer), some verb-based (decide_campagne, obtenir_cle_api), and the cao_* tools use a prefix convention. Still readable but not uniform.
41 tools is excessive for most servers, even one covering real estate and CAD. While the breadth is justified by multiple subdomains, the sheer number makes the surface hard to navigate and suggests possible over-fragmentation.
The tool set covers a wide range: rental lifecycle (annonce, EDL, quittance, relance), owner reporting, artisan documents, fiscal synthesis, CAD generation and verification, outreach, and guides. Minor gaps exist (e.g., no direct tenant/property CRUD), but for a document-generation server it is quite complete.
