NACHA/ACH ファイルの解析と検証を行う MCP (Model Context Protocol) サーバー。
Node.js 18 以降
npm
Related MCP server: @validatefin/mcp
git clone https://github.com/msuresh007/nacha-mcp.git
cd nacha-mcp
npm install
npm run buildこれにより src/ が dist/ にコンパイルされます。更新を取得した後は、npm run build を再実行してください。
計算上有効な小さなサンプルファイルが examples/sample.ach に含まれています。ビルド後、MCP クライアントなしでコマンドラインからどちらかのツールを直接実行して、すべてが正しく動作することを確認できます:
node -e "const {analyzeNachaFile}=require('./dist/analyze.js');const fs=require('fs');console.log(JSON.stringify(analyzeNachaFile(fs.readFileSync('examples/sample.ach','utf-8')),null,2))""valid": true と、空の issues 配列、1 つのバッチ、1 つのエントリが表示されるはずです。
parse_nacha_file — 指定したパスにある NACHA ファイルを完全な構造化 JSON にパースします: ファイルヘッダー、バッチ（ヘッダー、addenda 付きエントリ、コントロール）、ファイルコントロール。
summarize_nacha_file — NACHA ファイルをパースし、簡潔な要約を返します: バッチ数、エントリ総数、借方/貸方の合計金額、存在する SEC コード、検証上の問題。
どちらのツールも入力として file_path（ディスク上のファイルへの絶対パス）を 1 つ受け取ります。
パースには以下が含まれます:
構造検証（レコード長、レコードタイプの順序、バッチヘッダー/コントロールの対応付け）。
算術検証 — エントリハッシュ、借方/貸方の合計額、エントリ/addenda 件数は実際のエントリから再計算され、宣言されたバッチコントロールおよびファイルコントロールの値と照合されます。不一致は黙って受け入れられるのではなく、検証上の問題として報告されます。
IAT（International ACH Transaction）バッチと addenda タイプコード 10-18。
npm startサーバーは MCP プロトコルを使用して stdio 経由で通信します。単独で対話的に実行することを想定していません。下記のように MCP クライアント経由で使用してください。
プロジェクトディレクトリで、ビルド後:
claude mcp add nacha -- node "$(pwd)/dist/index.js"(Windows PowerShell では: claude mcp add nacha -- node "$PWD\dist\index.js")
使用するプロジェクトに .vscode/mcp.json を作成（既存の場合は追加）し、パスをこのリポジトリのクローンの絶対パスに置き換えます:
{
"servers": {
"nacha": {
"type": "stdio",
"command": "node",
"args": ["/absolute/path/to/nacha-mcp/dist/index.js"]
}
}
}または、コマンドパレット (Ctrl+Shift+P / ⇧⌘P) から MCP: Add Server を実行し、 Workspace を選択して、node を同じ引数で指定します。VS Code が同じ .vscode/mcp.json を書き込んでくれます。
保存後、Copilot Chat を開いて Agent モードに切り替えると、parse_nacha_file / summarize_nacha_file ツールが利用可能になります（VS Code は必要に応じてサーバを起動します）。
クライアントの MCP サーバー設定（例: claude_desktop_config.json）にエントリを追加し、パスをこのリポジトリのクローンの絶対パスに置き換えます:
{
"mcpServers": {
"nacha": {
"command": "node",
"args": ["/absolute/path/to/nacha-mcp/dist/index.js"]
}
}
}設定を追加した後、クライアントを再起動します。
src/constants.ts — SEC コード、トランザクションコード、ISO 国名/通貨コードの参照。
src/format.ts — 金額/日付/時刻フィールドのパース。
src/parser.ts — 構造的検証とレコード解析（ファイル/バッチ/エントリ/addenda）。
src/validate.ts — 算術的検証（ハッシュ、合計、件数）。
src/analyze.ts — パースと算術的検証の統合。
src/summarize.ts — 簡潔な要約ビュー。
src/index.ts — MCP サーバーとツールの登録。
ユニットテストは意図的に含まれていません。小さく、監査しやすいサーバーであり続けることを目的としています。
MIT — LICENSE を参照してください。
Parses a NACHA (ACH) fixed-width file into structured JSON: file header, batches (header, entries with addenda, control), and file control. Validates record structure and cross-checks entry hashes, debit/credit totals, and record counts, reporting any mismatches as issues.
| Name | Required | Description | Default |
|---|---|---|---|
| file_path | Yes | Absolute path to the NACHA/ACH file on disk. |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations are provided, so the description carries the burden. It discloses validation behavior well: cross-checking record counts, hashes, and debit/credit totals, and reporting mismatches as 'issues'. It could add a bit more about error behavior for unreadable or malformed files, but the existing detail is substantial.
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?
Two tight sentences, each carrying essential information: the first defines the output structure, the second the validation behavior. It is front-loaded with the core action and resource. No filler or repetition.
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?
Given one parameter, no annotations, and no output schema, the description is remarkably complete: it lists the CSV JSON composition, the validation checks, and the reporting of issues. An agent can confidently call the tool knowing what input to provide and what to do with the result.
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?
There is only one parameter and its schema description is 100% complete ('Absolute path to the NACHA/ACH file on disk'). The tool description does not add anything about the parameter beyond the schema, but with full schema coverage the baseline of 3 is appropriate.
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?
States a specific verb ('Parses') and resource ('NACHA/ACH fixed-width file') and names the output structure explicitly (file header, batches, control). It also distinguishes itself from sibling summarize_nacha_file by emphasizing structured JSON extraction plus validation rather than summarization.
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?
The description conveys the natural use case—get structured JSON from a NACHA file—but it does not explicitly contrast with summarize_nacha_file or state when the agent should choose parsing over summarizing. Usage context is implied clearly by the verb and output, but there is no explicit when/when-not guidance.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Parses a NACHA (ACH) file and returns a condensed summary: batch count, total entries, total debit/credit amounts, SEC codes present, and validation issues, without dumping every field of every record.
| Name | Required | Description | Default |
|---|---|---|---|
| file_path | Yes | Absolute path to the NACHA/ACH file on disk. |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
With no annotations provided, the description carries the full safety burden. It clearly states it parses and returns a summary, implying a read-only operation, and notes that it reports validation issues. However, it does not explicitly state side-effect-free behavior or potential error conditions on malformed files.
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?
A single, tightly packed sentence that front-loads the verb and resource, then enumerates specific output elements. No wasted words, and the critical exclusion is stated compactly.
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 simple one-parameter tool with no output schema, the description adequately specifies its return content with field-level detail. It falls just short of being fully complete by not describing the result structure or behavior on invalid files, though those gaps are minor.
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 coverage is 100% and the single file_path parameter is well-documented in the schema. The description does not add meaning beyond the schema, which is fine at the baseline of 3.
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?
The description states a specific action ('Parses... and returns') and explicitly lists the summary contents: batch count, total entries, total debit/credit amounts, SEC codes, and validation issues. The phrase 'without dumping every field of every record' clearly differentiates it from the sibling parse_nacha_file.
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?
The description explicitly frames when to use this tool: when a condensed summary is desired rather than full record-level detail. The contrast 'without dumping every field' effectively communicates the when-not and implicitly routes to the sibling for detailed extraction.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
v0.1.0parse_nacha_filesummarize_nacha_fileScored across 2 tools
The two tools have clearly different outputs—one returns the full parsed structure and one returns a condensed summary—so an agent can pick based on response size. There is mild overlap because both parse the file, but the naming and descriptions sufficiently differentiate their purposes.
Both tool names follow the same verb_noun_file pattern: parse_nacha_file and summarize_nacha_file. The naming is predictable, consistent, and accurately reflects the intent of each tool.
With only two tools, the server is intentionally narrow, but both tools serve a clear and distinct purpose for NACHA file handling. The count is slightly thin but not inappropriate given the focused read-only scope.
For a server focused on reading and validating NACHA files, the surface covers the essential workflows: full detailed parsing and condensed summarization. It could benefit from additional capabilities like file creation or export, but those seem outside its apparent scope.
Parse WebVTT, SRT, or TTML for conformance, timing, overlaps, line length, and reading speed.
Turn bank statement PDFs into categorized, balance-checked transactions and reports.
Turn bank statement PDFs, CSVs, XLSX and OFX into categorised transactions plus a summary.
Validate HMDA LAR files against CFPB edit checks for community banks and credit unions.