Analyse Pièce Comptable for Openclaw

An intelligent read-only analysis skill designed to extract financial metrics, identify compliance issues, and provide professional accounting summaries for individual documents.

trendex
v1.0.0
Jun 4, 2026
0
529
0

Install & Download

1. ClawHub CLI

The fastest way to install a skill directly from the registry.

npx clawhub@latest install analyse-piece-comptable

2. Manual Installation

Copy the skill folder to one of these locations

Global
~/.openclaw/skills/
Workspace
<project>/skills/

Priority: Workspace > Local > Bundled

3. Prompt Installation

Copy this prompt to OpenClaw to install it automatically.

Help me install analyse-piece-comptable using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).

Prefer to download?

Get the raw skill files in a ZIP archive.

What is Analyse Pièce Comptable?

The analyse-piece-comptable skill is an essential tool tailored for financial workflows within Openclaw Skills. Acting strictly in a read-only capacity, it enables developers and accounting agents to quickly inspect individual financial documents—such as invoices, bank statements, or expense receipts—without modifying, moving, renaming, or archiving the source files.

By processing documents through a deterministic parsing pipeline, the skill extracts critical monetary data, tax evaluations, transaction dates, and stakeholder identities. It then performs multi-point consistency checks to detect anomalies like incorrect calculations, fraudulent dates, or malformed identifiers, turning raw document data into structured insight.

Analyse Pièce Comptable Use Cases

  • Evaluating a single invoice to verify VAT (TVA) accuracy and total amounts before ledger entry.
  • Scanning a bank statement to validate closing balance calculations and verify IBAN parameters.
  • Reviewing an isolated expense report (note de frais) to ensure matching employer-to-employee reimbursement values.
  • Pre-screening an newly arrived email attachment for accounting anomalies before any formal classification.

How Analyse Pièce Comptable Works

  1. The accountant or automated workflow triggers the skill by passing an isolated financial document to the underlying script.
  2. The system invokes scripts/analyse.py along with a reference date parameter to evaluate the document deterministically.
  3. The script extracts critical fields and runs programmatic consistency evaluations to build an error-and-alert metadata array.
  4. A structured JSON payload containing document taxonomy, confidence scores, extracted fields, and check categories is returned to stdout.
  5. The AI agent translates this raw JSON output into a natural, professional accounting summary, skipping all technical jargon.

Analyse Pièce Comptable Setup

To deploy and use this analysis skill, ensure your local workspace contains the required python environment and dependencies. Run the analysis script against an isolated document using the command-line interface:

python3 scripts/analyse.py <path_to_document_file> --date-ref <YYYY-MM-DD>

The --date-ref parameter must match the current date to accurately spot post-dated entries or downstream anomalies.

Analyse Pièce Comptable Data Schema & Taxonomy

The underlying analyzer evaluates text layers to populate a standardized schema. When text_found returns false, it signals a scanned image or photo requiring visual fallback handling.

Output Fields Taxonomy

Key Type Description
kind String Type of document: invoice, bank-statement, note-de-frais, or other
text_found Boolean True if a text layer is present; False if it is a flat scan or photo
confidence String Extraction certainty index: haute, moyenne, or faible
fields Object Extracted metadata dictionary containing IDs, dates, totals, VAT, and parties
checks Array Collection of anomaly objects containing validation codes, severity, and details

Integrity Validation Codes

  • champ_manquant: Essential tracking fields missing from the document layout.
  • tva_incoherente: Stated VAT deviates by more than 5% from standard rate calculations.
  • montant_aberrant: General math failure where Net + VAT does not equal Gross Total.
  • date_future: Issue date is set to an impossible future point relative to reference date.
  • echeance_anterieure: Payment deadline falls before the initial issue date.
  • iban_invalide: Bank statement routing digits fail standard checksum key validation.
  • solde_incoherent: Statement opening balance combined with transaction drift does not match closing balance.
  • releve_non_parseable: Bank statement text format prevents clean structural parsing.
  • sans_texte: Document lacks an accessible text layer.

Analyse Pièce Comptable Advanced Features

  • Deterministic Fallback Pipeline: Automatically routes to an eye-balled, human-in-the-loop styled visual parsing workflow with built-in warnings when digital text structures are missing.
  • Multi-Severity Check Hierarchy: Categorizes validation discrepancies into actionable error tiers: bloquant (critical/unreliable), alerte (requires human review), and info (minor notation).
  • Business-Language Translation Layer: Formats complex programmatic JSON payloads into crisp, professional French summaries that skip script details or system flags, optimizing the interface for trained accountants.
  • Scope Isolation Guardrails: Enforces a strict read-only design, preventing accidental data modification and allowing this capability to be safe for deployment alongside specialized sorting modules in Openclaw Skills ecosystems.

SKILL.md


Loading

Related Openclaw Skills

METADATA

Github Stars: 0
forks: 0

Featured*