Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
git clone --depth 1 https://github.com/michielinksee/bantouWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/commands/michielinksee/bantou/nightly)<a href="https://agentmods.dev/commands/michielinksee/bantou/nightly"><img src="https://agentmods.dev/badge/commands/michielinksee/bantou/nightly/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/commands/michielinksee/bantou/nightly"><img src="https://agentmods.dev/badge/commands/michielinksee/bantou/nightly.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00015 | $0.00486 |
| Opus 5 | $0.00008 | $0.00243 |
| Sonnet 5 | $0.00003 | $0.00097 |
| Haiku 4.5 | $0.00002 | $0.00049 |
Grade A, and why
nightly scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 9d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
What it actually says
/bantou:nightly
Trigger the nightly batch processing run for all connected client companies.
Usage
/jp-accounting:nightly
No arguments required. Processes all companies connected via freee API.
What happens
- Enumerate companies — Lists all companies accessible via the configured freee access token
- Fetch transactions — For each company, fetches unprocessed transactions since the last run
- Classify — Runs the full 3-stage pipeline on every transaction:
- Stage 0: Exclusion filter
- Stage 1: Memory recall + keyword matching
- Stage 2: AI classification (fallback)
- Route by confidence — Sorts results into tiers:
- High (>= 85%): auto-classified, no review needed
- Medium (70-84%): auto-classified, flagged for morning review
- Low (< 70%): held for human review
- Generate summary — Produces a batch run summary
Output
A summary is displayed after processing:
夜間バッチ処理完了
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
処理企業数: 12
総取引数: 847
除外: 63
自動分類 (高): 612
自動分類 (中): 108
要確認 (低): 64
Memory ヒット率: 38.2%
API トークン使用: ~12,400
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
Low-confidence items are listed with details for morning review.
Scheduling
For automated nightly execution, configure via Anthropic Routines or the system scheduler. Recommended schedule: 02:00 JST weekdays.
Prerequisites
FREEE_ACCESS_TOKENmust be configured in the plugin settings- At least one freee company must be connected
- For non-freee clients, use
/jp-accounting:importinstead
Notes
- The batch is idempotent; re-running does not create duplicates
- If a company fails, processing continues for remaining companies
- Run logs are stored locally for audit trail
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 9d ago First seen · 70 lines · 15 tokens per session scan A b6cee47a3c41
nightly is a command published in the GitHub repository michielinksee/bantou (2 stars, last pushed 18d ago), licensed MIT. It adds 15 tokens to every session and 486 once invoked, about $0.0001 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other commands, from other repositories
ingest
Ingest financial data (bank/credit-card statements, CSV exports, receipts, or a plain-text description) into the book as validated double-entry transactions.
init
Initialise Countbean — connect to your hosted book (approving in the browser) or create a local one, plus a starter chart of accounts.
assess
Review the book and report what the numbers actually show — run rate, categories, changes, unusual entries, and gaps in the data.
connect
Connect Claude to your hosted Countbean book — approve it in your browser, or paste a key you already have.
report
Generate a financial report from the book — a styled HTML page and/or an Excel workbook (balance sheet, income statement, transactions).
status
Show the current state of your Countbean book — balances, net worth, account count, and recent changes.