A monthly report of observed refusals, showing changes from the previous month and which tools or tables caused the most blocks. It also compares these findings with the reconciliation backlog.

Use it to review monthly refusal trends, compare months, rank problem areas, and check whether missing refusals indicate stale verification data.

It helps distinguish worsening data quality from ordinary backlog noise and identify where the latest problems began.

verbale — Command for Claude Code by capazme/mcp-legal-it

Install

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.

Clone the repo

git clone --depth 1 https://github.com/capazme/mcp-legal-it

Made for: Claude Code.

Or install legal-it, the plugin that ships this one along with the rest of its 23 skills, 10 commands, 6 agents, 1 hook, 1 MCP server.

Wrote 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.

README.md

[![agentmods](https://agentmods.dev/badge/commands/capazme/mcp-legal-it/verbale/github.svg)](https://agentmods.dev/commands/capazme/mcp-legal-it/verbale)

Your own site

<a href="https://agentmods.dev/commands/capazme/mcp-legal-it/verbale"><img src="https://agentmods.dev/badge/commands/capazme/mcp-legal-it/verbale/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.

Your own site · 80×15

<a href="https://agentmods.dev/commands/capazme/mcp-legal-it/verbale"><img src="https://agentmods.dev/badge/commands/capazme/mcp-legal-it/verbale.svg" alt="Reviewed on agentmods" width="80" height="20"></a>

Per session 31 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.

When invoked 639 The whole file, excluding the scripts and references it only reads on demand.

Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.

Origin original No closer match found in the catalogue.

Token cost

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	$0.00031	$0.00639
Opus 5.5	$0.00012	$0.00256
Sonnet 5.5	$0.00006	$0.00128
Haiku 4.5	$0.00003	$0.00064

Measured 26d ago against content hash 150818171af9, method: parsed. Prices are Anthropic first-party input rates as of 2026-10-07, from the pricing page.

Security

Grade A, and why

verbale 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 26d 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.

plugin/commands/verbale.md · 47 lines

What it actually says

Comando /verbale — la serie storica dei rifiuti

Chiama legal-it:verbale_mensile (con months: $ARGUMENTS se indicato) e mostra prima il grafico della serie (il campo grafico: una barra per mese, scala sul massimo della finestra, il mese corrente marcato) — è il confronto fra i mesi, a colpo d'occhio. Poi la tabella: una riga per mese con rifiuti, accettazioni, il delta col mese precedente e il tool/tabella che ha guidato il mese.

Cosa leggere nella serie

Il mese corrente: se delta_mese_precedente è positivo, il primo posto della classifica (top_tool, top_tabella) dice cosa è peggiorato — parti da lì, non dal backlog completo.
Il confronto col backlog: chiama anche legal-it:backlog_riconciliazione e incrocia: una tabella con rifiuti osservati ma zero rifiutano statici indica che la qualità dei dati è peggiorata dopo l'ultima verifica (o che qualcuno ha rimosso una alternativa); una tabella in cima al backlog senza rifiuti osservati è rumore, non priorità.
Un mese a zero non è un buco: o il verbale è spento (disponibile: false — dillo esplicitamente e suggerisci LEGAL_REFUSAL_LEDGER=on), o il mese è andato davvero liscio.

Manutenzione senza aprire il codice

Stessa aggregazione da terminale: python3 scripts/verbale-report.py (--out report.md per archiviarlo, --strict per far fallire una pipeline quando il mese è peggiorato).
Il verbale registra solo rifiuti e accettazioni di precisione — mai dati di causa. Nessun regime di riservatezza necessario.
Dopo una riconciliazione (vedi /dati), il mese successivo deve mostrare il calo: se non lo mostra, la riconciliazione non ha toccato i tool che bloccano davvero.

Nota

Il report legge <MCP_CACHE_DIR>/refusals.jsonl, lo stesso file che il middleware scrive a ogni rifiuto/accettazione: tool e script non possono mai disaccordarsi sui numeri.

Changes

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.

26d ago First seen · 47 lines · 31 tokens per session scan A 150818171af9

Subscribe to this mod's changes

verbale is a command published in the GitHub repository capazme/mcp-legal-it (24 stars, last pushed 2d ago), licensed Apache-2.0. It adds 31 tokens to every session and 639 once invoked, about $0.0001 per session on Opus 5.5. 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 2026-09-22.

Analytics and metrics LLM Legal and compliance claude italian-law legal-tech mcp model-context-protocol

check

View the latest screenshot from /Pictures/Screenshots.

jmagar/claude-homelab · 10 tokens

gsd:discuss-phase

Gather phase context through adaptive questioning before planning.

mrboups/xbrain · 16 tokens

gsd:explore

Socratic ideation and idea routing — think through ideas before committing to plans.

mrboups/xbrain · 21 tokens

eu-cloud-sovereignty

You are helping an enterprise architect generate an EU Cloud Sovereignty Framework Assessment for an organisation procuring, specifying, or evaluating cloud services against the European Commission's EU Cloud Sovereignty Framework (v1.2.1, October 2025). The framework supplements security assurance (SecNumCloud, EUCS…

tractorjuice/arckit-claude · 34 tokens

at-bvergg

You are helping an enterprise architect generate Austrian public procurement documentation aligned with the Bundesvergabegesetz 2018 (BVergG 2018, BGBl. I Nr. 65/2018), as amended most recently by the Vergaberechtsgesetz 2026 (BGBl. I Nr. 8/2026 — the largest procurement reform since 2018; in force 1 March 2026, with…

tractorjuice/arckit-claude · 50 tokens

atrs

Generate Algorithmic Transparency Recording Standard (ATRS) record for AI/algorithmic tools.

tractorjuice/arckit-claude · 17 tokens