An end-to-end accounting document processing orchestrator that automates sorting, analyzing, identifying, and filing financial documents.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install pipeline-comptable
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install pipeline-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 pipeline-comptable?
The pipeline-comptable skill acts as an intelligent orchestrator for incoming accounting documents, mirroring the workflow of a human assistant handling morning mail. Built for Openclaw Skills, it coordinates specialized single-task sub-modules to guide documents effortlessly from raw entry points down to final storage and payment reconciliation.
Rather than embedding complex processing logic directly within itself, this skill functions strictly as an engine that sequences four dedicated micro-brakes: incoming mail sorting, deep document analysis, client identification, and document filing. This architectural separation ensures that if an underlying analysis or sorting engine updates, the main orchestration pipeline continues to function perfectly without code changes.
pipeline-comptable Use Cases
- End-to-End Email Inbox Automation: Processing incoming unread emails with attached invoices, receipts, or bank statements automatically.
- Bulk Batch Document Processing: Handling folders of mixed, pre-downloaded accounting documents from a manual drop zone.
- Multi-Client Document Categorization: Automatically identifying which client or supplier an incoming document belongs to and deciding whether it represents a purchase or a sale.
- Automated Ledger Filing & Reconciliation: Sorting documents into their designated folder paths and triggering payment matching sequences automatically.
How pipeline-comptable Works
- Tri Courrier Entrant (Step 1): Filters incoming emails to determine if they contain accounting relevance and extracts valid document attachments.
- Analyse Pièce Comptable (Step 2): Parses the extracted document to extract critical metadata, including document type, explicit dates, and transaction amounts.
- Identification Client (Step 3): Matches the analyzed data against known customer records to identify the client identity and classify the ledger category (purchase vs. sale).
- Classement Document (Step 4): Stores the file in its exact, structured target folder within the directory tree.
- Rapprochement Paiements (Step 5): Automatically kicks off a reconciliation pass to cross-reference bank statements with filed invoices.
pipeline-comptable Setup
The pipeline-comptable script accepts inputs from raw email JSON buffers or local storage directories. Configure paths and execute the pipeline using standard shell arguments:
# Run the full pipeline starting from a raw incoming email file
python3 scripts/pipeline.py --email <mail.json> --clients-root <path/to/clients>
# Run the processing pipeline starting from a local folder of pre-downloaded documents
python3 scripts/pipeline.py --inbox <folder_path> --clients-root <path/to/clients>
pipeline-comptable Data Schema & Taxonomy
The skill manages processing state by passing a unified JSON document object down the line. It produces two distinct outputs:
| Output Destination | Data Format / Content | Purpose |
|---|---|---|
Standard Output (stdout) |
Consolidated JSON Report | A complete, machine-readable array containing detailed objects for every processed document file, populated by each sequence step. |
Standard Error (stderr) |
Human-Readable Log | A clean text summaries highlighting filed items, detected duplicates, items pending validation, and explicit accounting queries. |
Output Constraints
To ensure clean reporting, technical paths, low-level execution data, script filenames, or explicit references to 'JSON' or 'index' are strictly stripped from human-facing error logs.
pipeline-comptable Advanced Features
- No-Reconciliation Flag: Run the entire processing stream while omitting the resource-heavy payment matching phase by appending the
--no-rapprochementflag. - Decoupled Architecture Support: Utilize individual sub-modules independently for single-file tasks, allowing isolated upgrades to separate Openclaw Skills blocks without breaking the overall process.
- Human-in-the-Loop Conflict Resolution: Automatically identifies ambiguous data points or incomplete records and aggregates precise questions for the accountant rather than guessing.
SKILL.md
Loading
Related Openclaw Skills
zdat-publish-skill
An advanced content distribution and scheduling agent skill that reformats, humanizes, and cross-publishes articles across multiple platforms including WeChat, Zhihu, Toutiao, and Xiaohongshu.
ZDAT Crawl Skill
An AI-powered intelligence harvesting and threat crawling skill designed to monitor zero-defect topics, track multi-platform sentiments, and trigger real-time alerts.
ZDAT Chat Skill
ZDAT Chat Skill automates social media comment monitoring, matches user intent with intelligent scripts, logs high-intent leads into a local ledger, and implements anti-rate-limiting measures.
e-facture-rapprochement
An automated billing and bank reconciliation engine that normalizes multi-format invoices and bank statements using local scripts and LLM vision extraction.
Goofish Search
Goofish Search is an automated browser skill designed to search Xianyu/Goofish for secondhand products, apply strict merchant filtering, sort by price, and output optimized top results.
DashClaw Platform Intelligence
A high-performance AI agent skill designed for live system querying, API mapping, and comprehensive environment governance within the DashClaw platform.
Featured*
