Mcp Beancount — independently scanned and version-tracked by SaferSkills.
SaferSkills independently audited Mcp Beancount (Agent Skill) and scored it 100/100 (green). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 0 high-severity and 0 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.
Findings & checks · 0 flagged
Every scanned point with the score it earned and what moved between them.
First recorded scan — no prior version to compare against.
The primary manifest — the file an agent reads to learn what this artifact does.
Read-only MCP server that gives AI agents structured access to a Beancount personal finance ledger.
uv sync| Variable | Required | Default | Description |
|---|---|---|---|
BEANCOUNT_FILE | ✅ | — | Absolute path to your .beancount ledger file |
ACCOUNT_ALLOWLIST | ❌ | (none) | Comma-separated account prefix whitelist (e.g. Assets:Bank,Expenses:) |
BEANCOUNT_RELOAD | ❌ | false | Reload ledger on every tool call (development mode) |
BASE_CURRENCY | ❌ | operating_currency from ledger | Override the base currency for net_worth_converted (e.g. CHF) |
Copy .env.example to .env and fill in your values:
cp .env.example .envBEANCOUNT_FILE=/path/to/ledger.beancount uv run mcp-beancountOr with a .env file (requires dotenv or equivalent):
export BEANCOUNT_FILE=/path/to/ledger.beancount
uv run mcp-beancountAdd to ~/Library/Application Support/Claude/claude_desktop_config.json:
{
"mcpServers": {
"beancount": {
"command": "uv",
"args": ["run", "--project", "/path/to/mcp-beancount", "mcp-beancount"],
"env": {
"BEANCOUNT_FILE": "/absolute/path/to/ledger.beancount",
"ACCOUNT_ALLOWLIST": ""
}
}
}
}Add to your OpenClaw MCP config:
{
"mcpServers": {
"beancount": {
"command": "uv",
"args": ["run", "--project", "/path/to/mcp-beancount", "mcp-beancount"],
"env": {
"BEANCOUNT_FILE": "/absolute/path/to/ledger.beancount"
}
}
}
}| Tool | Signature | Description |
|---|---|---|
get_net_worth | (date?: str) | Net worth (assets − liabilities) as of date, multi-currency with price conversion |
get_balances | (account_pattern: str) | Non-zero balances matching account prefix/glob |
get_income_statement | (year: int, month?: int) | Income vs expenses for a period |
get_transactions | (account?: str, since?: str, limit?: int) | Recent transactions, filterable |
query | (bql: str) | Raw BQL query (read-only) |
# Net worth today (multi-currency, per-currency breakdown + converted total)
get_net_worth()
# Returns:
# {
# "as_of": "2026-03-28",
# "base_currency": "CHF",
# "assets": {
# "Assets:Bank:UBS": {"CHF": 50000.0},
# "Assets:Broker:IBKR": {"USD": 30000.0}
# },
# "liabilities": {
# "Liabilities:CreditCard": {"CHF": -2000.0}
# },
# "total_assets": {"CHF": 50000.0, "USD": 30000.0},
# "total_liabilities": {"CHF": -2000.0},
# "net_worth": {"CHF": 48000.0, "USD": 30000.0},
# "net_worth_converted": 74700.0, # CHF 48000 + USD 30000 * 0.89
# "skipped_positions": [] # currencies with no price directive
# }
# Net worth as of 2025-12-31
get_net_worth(date="2025-12-31")
# All bank account balances
get_balances(account_pattern="Assets:Bank:")
# 2026 income statement
get_income_statement(year=2026)
# Q1 2026 income statement
get_income_statement(year=2026, month=1)
# Last 20 food expenses
get_transactions(account="Expenses:Food", limit=20)
# Transactions since 2026-01-01
get_transactions(since="2026-01-01")
# Raw BQL query
query("SELECT account, sum(position) WHERE account ~ 'Assets' GROUP BY account")BEANCOUNT_FILE=/path/to/any.beancount uv run pytestOr using the included sample.beancount (tests use this automatically via fixtures):
uv run pytestBEANCOUNT_FILE is set at startup; never exposed through tool argumentsACCOUNT_ALLOWLIST restricts all tools to specific account prefixes~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.