pacs008-mcp

pacs008-mcp

An MCP server that exposes the pacs008 ISO 20022 FI-to-FI Customer Credit Transfer library as tools for AI agents and assistants, enabling generation, validation, and parsing of pacs.008 credit transfer XML messages.

Category
访问服务器

README

pacs008-mcp: An MCP Server for ISO 20022 FI-to-FI Credit Transfers

pacs008-mcp banner

PyPI Version Python Versions PyPI Downloads Licence Tests Quality Documentation

A Model Context Protocol server that exposes the pacs008 ISO 20022 FI-to-FI Customer Credit Transfer library as tools for AI agents and assistants — discover message types and scheme profiles, validate records against the JSON Schema and against a rail's usage guidelines, generate validated XML, validate raw XML against the bundled XSD, and parse inbound messages, all from your favourite MCP client.

Latest release: v0.0.4 — 15 MCP tools over stdio, all backed by the pacs008 library, for Python 3.10+. Adds convert_mt103, the legacy SWIFT MT103 → pacs.008 (MT→MX) migration path.

Contents

Overview

The Model Context Protocol (MCP) is an open standard that lets AI agents and assistants discover and call external tools in a uniform way. pacs008-mcp is an MCP server that turns the pacs008 library into a set of first-class agent tools, so an assistant can generate, validate, and parse ISO 20022 pacs.008 FI-to-FI Customer Credit Transfer XML messages — and the related pacs.002/.004 status and return messages — directly from a conversation.

Every tool is a thin, typed wrapper over the pacs008 library — the same package used by the CLI and REST API — so all interfaces behave identically. Tools return JSON-serialisable data; on a validation error they return an {"error": ...} payload rather than raising.

flowchart LR
    A["MCP client<br/>(Claude Desktop, IDE, agent)"] -->|stdio| B["pacs008-mcp"]
    B -->|delegates to| C["pacs008 library"]
    C -->|render + validate| D["ISO 20022 pacs XML"]

Install

pacs008-mcp runs on macOS, Linux, and Windows and requires Python 3.10+ and pip. It pulls in the core pacs008 library and the MCP SDK automatically.

python -m pip install pacs008-mcp

Note: while the core pacs008 library is not yet on PyPI, install it from source first:

python -m pip install "git+https://github.com/sebastienrousseau/pacs008.git"
python -m pip install pacs008-mcp

Quick Start

Launch the server over stdio (the FastMCP default transport):

pacs008-mcp

Register it with any MCP client (e.g. Claude Desktop) by adding it to the client's configuration:

{
  "mcpServers": {
    "pacs008": { "command": "pacs008-mcp" }
  }
}

Tools

All tools wrap the pacs008 library, so they behave identically to the CLI and REST API.

Tool Purpose
list_message_types List the 20 supported ISO 20022 pacs message types
list_schemes List the registered scheme / usage-guideline profiles
get_scheme Inspect a scheme profile's rules
get_required_fields Required input fields for a message type
get_input_schema Full input JSON Schema for a message type
validate_records Validate flat records against a message type's schema
validate_scheme Validate records against a scheme's usage guidelines
generate_message Generate a validated pacs XML message
validate_xml Validate a raw XML string against the bundled XSD
parse_message Parse & classify an inbound ISO 20022 message
convert_mt103 Convert a legacy SWIFT MT103 into pacs.008-ready records (MT→MX migration)
classify_address Classify a postal address as structured / hybrid / unstructured
validate_address Validate one postal address against an address policy
repair_address Upgrade legacy unstructured address lines toward hybrid/structured form
validate_addresses Batch-validate every party address across payment rows

November 2026 structured-address cliff

On 14 November 2026, fully unstructured postal addresses are decommissioned across SWIFT CBPR+, HVPS+, TARGET2 RTGS, CHAPS, Fedwire and Lynx — after that date, any cross-border or high-value payment carrying an unstructured-only postal address is rejected at the rail. The four address tools above wrap the pacs008 library's standards.address module so an agent can get ahead of the deadline: classify_address shows where an address stands, validate_address / validate_addresses enforce a policy (defaulting to the cliff rule hybrid_or_structured, which rejects unstructured addresses), and repair_address runs country-aware heuristics (GB, US, DE, FR, JP, plus a best-effort fallback) to lift legacy address lines into hybrid form. The repair step is experimental — audit its output before submitting downstream.

Using the tools

You can invoke the tools in-process — without a transport — straight through the FastMCP instance. This mirrors what an agent receives over stdio. The runnable version of this snippet lives in examples/mcp_tools.py.

import asyncio

from pacs008_mcp.server import server

record = [
    {
        "msg_id": "MSG001",
        "creation_date_time": "2026-01-15T10:30:00",
        "nb_of_txs": "1",
        "settlement_method": "CLRG",
        "interbank_settlement_date": "2026-01-15",
        "end_to_end_id": "E2E001",
        "interbank_settlement_amount": "1000.00",
        "interbank_settlement_currency": "EUR",
        "charge_bearer": "SHAR",
        "debtor_name": "Debtor Corp",
        "debtor_agent_bic": "DEUTDEFF",
        "creditor_agent_bic": "COBADEFF",
        "creditor_name": "Creditor Ltd",
    }
]


async def main() -> None:
    async def call(name, args):
        result = await server.call_tool(name, args)
        content = result[0] if isinstance(result, tuple) else result
        return content[0].text if content else ""

    print(await call("list_schemes", {}))
    xml = await call("generate_message",
                     {"message_type": "pacs.008.001.08", "records": record})
    print(xml[:46])  # -> <?xml version="1.0" encoding="UTF-8"?> ...


asyncio.run(main())

Run it directly:

python examples/mcp_tools.py

Development

pacs008-mcp uses Poetry and mise.

git clone https://github.com/sebastienrousseau/pacs008-mcp.git && cd pacs008-mcp
mise install
poetry install
poetry shell

This package depends on the core pacs008 library. Until it is on PyPI, install it from source first: pip install "git+https://github.com/sebastienrousseau/pacs008.git".

A Makefile orchestrates the quality gates (kept in lockstep with CI):

make check        # all gates (REQUIRED before commit)
make test         # pytest
make lint         # ruff + black
make type-check   # mypy --strict

Related MCP Servers

Part of the ISO 20022 MCP Suite — open-source, Apache-2.0 licensed MCP servers for banking and financial-services AI agents:

Server Purpose
pain001-mcp Generate & validate ISO 20022 pain.001 payment files (v03–v12, pain.008, SEPA) with rulebook checks
camt053-mcp Parse & reconcile ISO 20022 camt.053 bank-to-customer statements — CBPR+/HVPS+ ready
acmt001-mcp Generate & validate ISO 20022 acmt account-management messages
bankstatementparser-mcp Parse bank statements (BAI2, MT940/MT942, CAMT.053, OFX, CSV) into structured transactions
noyalib-mcp Lossless YAML 1.2 parsing, formatting & validation (Rust, 100% spec compliance)

MCP Registry

mcp-name: io.github.sebastienrousseau/pacs008-mcp


Licence

Licensed under the Apache Licence, Version 2.0. Any contribution submitted for inclusion shall be licensed as above, without additional terms.

Contribution

Contributions are welcome — see the contributing instructions. Thanks to all contributors.

Acknowledgements

Built on the pacs008 ISO 20022 FI-to-FI Customer Credit Transfer library and the Model Context Protocol Python SDK.

推荐服务器

Baidu Map

Baidu Map

百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。

官方
精选
JavaScript
Playwright MCP Server

Playwright MCP Server

一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。

官方
精选
TypeScript
Magic Component Platform (MCP)

Magic Component Platform (MCP)

一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。

官方
精选
本地
TypeScript
Audiense Insights MCP Server

Audiense Insights MCP Server

通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。

官方
精选
本地
TypeScript
VeyraX

VeyraX

一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。

官方
精选
本地
graphlit-mcp-server

graphlit-mcp-server

模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。

官方
精选
TypeScript
Kagi MCP Server

Kagi MCP Server

一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。

官方
精选
Python
e2b-mcp-server

e2b-mcp-server

使用 MCP 通过 e2b 运行代码。

官方
精选
Neon MCP Server

Neon MCP Server

用于与 Neon 管理 API 和数据库交互的 MCP 服务器

官方
精选
Exa MCP Server

Exa MCP Server

模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。

官方
精选