← Back to Skills Marketplace
90
Downloads
0
Stars
0
Active Installs
2
Versions
Install in OpenClaw
/install nfs-e-parser
Description
NFS-e field extractor for Brazilian agents. 100% field accuracy on São Paulo NFS-e invoices (auxiliar-nfs-e + Surya). Extracts CNPJ, prestador, tomador, valo...
Usage Guidance
This skill appears to do what it advertises (São Paulo NFS-e field extraction), but it relies on pip-installing 'surya-ocr' and cloning a GitHub repository before executing code. Before installing or running it: (1) inspect the GitHub repo (Tlalvarez/Auxiliar-ai) and the specific parser.py to confirm no network exfiltration or unexpected behavior; (2) review the PyPI package 'surya-ocr' (owner, source, and release history) and avoid running it on sensitive documents until you trust the package; (3) run installs and tests inside an isolated virtualenv or sandbox container, not on a production host; (4) test the parser on synthetic or redacted invoices first; and (5) consider scanning cloned code with your usual security tools (static analysis, dependency checks). If you can provide the GitHub repo link (actual content) or the parser.py file, I can give a higher-confidence assessment.
Capability Analysis
Type: OpenClaw Skill
Name: nfs-e-parser
Version: v0.1.1
The skill bundle in SKILL.md provides instructions for an AI agent to extract data from Brazilian NFS-e invoices by performing high-risk shell operations, including pip install and git clone from a remote repository (github.com/Tlalvarez/Auxiliar-ai.git). While the behavior aligns with the stated purpose of invoice parsing, the reliance on fetching and executing unverified external code via shell commands constitutes a significant supply chain risk. No clear evidence of malicious intent was found, but the broad execution capabilities required make it suspicious.
Capability Assessment
Purpose & Capability
Name/description match the requested binaries and actions: Python and git are reasonable for a Python parser that clones a GitHub repo and runs OCR. No unrelated credentials, config paths, or unrelated system binaries are requested.
Instruction Scope
SKILL.md is instruction-only and stays on-task (create venv, pip install surya-ocr and transformers, git clone repo, run surya_ocr, parse OCR text, validate CNPJ). It does instruct reading local PDF folders and running CLI tools on them (expected for invoice parsing) but grants the agent broad discretion to run OCR on any files the agent can access — review which folders the agent will process and avoid pointing it at sensitive or unrelated documents without oversight.
Install Mechanism
There is no formal install spec in the registry; the runtime instructions tell users to pip install surya-ocr and transformers and to git clone https://github.com/Tlalvarez/Auxiliar-ai.git. GitHub is a normal host, but pip-installing third-party packages and executing cloned scripts means arbitrary code from PyPI/GitHub will run. This is proportional to a Python-based parser but carries the usual risk of running unreviewed packages.
Credentials
No environment variables, credentials, or config paths are requested. The skill does not ask for unrelated secrets or system-level tokens, which is appropriate for a local invoice parser.
Persistence & Privilege
The skill is not always-enabled and does not declare modifications to other skills or system-wide settings. The runtime workflow clones and copies a parser script and installs packages into a virtualenv — normal for a local Python tool and not an escalation of platform privileges.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install nfs-e-parser - After installation, invoke the skill by name or use
/nfs-e-parser - Provide required inputs per the skill's parameter spec and get structured output
Version History
vv0.1.1
- Clarifies that the JSON response example contains only fictitious data and well-known test CNPJs, not real client data.
- Updates documentation to improve privacy and structure illustration, using representative fake invoice values.
- Generalizes example paths and workflow steps for broader use cases.
- No changes to code or dependencies; documentation improvements only.
vv0.1.0
Initial release of the nfs-e-parser skill.
- Extracts structured fields from São Paulo NFS-e (service invoice) PDFs with 100% tested accuracy on published Surya OCR corpus.
- Outputs JSON with key tax and invoice fields (CNPJ, prestador, tomador, valor_servico, código_serviço, ISS, retenções, etc.).
- Includes CNPJ validation and field completeness checks for bookkeeping and accountant handoff.
- Designed for use with Surya OCR; instructions provided for setup and batch processing.
- Limitations: supports only São Paulo layout; other municipalities not yet covered.
Metadata
Frequently Asked Questions
What is NFS-e Parser — Brazilian invoice field extraction?
NFS-e field extractor for Brazilian agents. 100% field accuracy on São Paulo NFS-e invoices (auxiliar-nfs-e + Surya). Extracts CNPJ, prestador, tomador, valo... It is an AI Agent Skill for Claude Code / OpenClaw, with 90 downloads so far.
How do I install NFS-e Parser — Brazilian invoice field extraction?
Run "/install nfs-e-parser" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is NFS-e Parser — Brazilian invoice field extraction free?
Yes, NFS-e Parser — Brazilian invoice field extraction is completely free, licensed under MIT-0. You can download, install and use it at no cost.
Which platforms does NFS-e Parser — Brazilian invoice field extraction support?
NFS-e Parser — Brazilian invoice field extraction is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created NFS-e Parser — Brazilian invoice field extraction?
It is built and maintained by Tlalvarez (@tlalvarez); the current version is vv0.1.1.
More Skills