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.
npx skills add Synthos-Logic/penalista-italia --skill penalista-archiviogit clone --depth 1 https://github.com/Synthos-Logic/penalista-italiaWrote 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/skills/synthos-logic/penalista-italia/penalista-archivio)<a href="https://agentmods.dev/skills/synthos-logic/penalista-italia/penalista-archivio"><img src="https://agentmods.dev/badge/skills/synthos-logic/penalista-italia/penalista-archivio/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/skills/synthos-logic/penalista-italia/penalista-archivio"><img src="https://agentmods.dev/badge/skills/synthos-logic/penalista-italia/penalista-archivio.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 | $0.00228 | $0.01933 |
| Opus 5 | $0.00114 | $0.00966 |
| Sonnet 5 | $0.00046 | $0.00387 |
| Haiku 4.5 | $0.00023 | $0.00193 |
Grade A, and why
penalista-archivio 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 12d 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.
How it starts
The opening of the file, as written. The whole thing — 119 lines — stays where its author put it; the contents beside it link to each section on GitHub.
penalista-archivio — Conversione e indicizzazione della Knowledge Base
Cosa fa
Trasforma documenti (in primis PDF a testo nativo: rassegne del Massimario, sentenze, atti, dottrina) in Markdown pulito con marcatori di pagina e genera un indice a due livelli che rende la KB interrogabile a basso costo di token e — soprattutto — ancorabile: ogni citazione potrà essere ricondotta a un file e a una pagina reali.
Due ragioni per cui esiste:
- Economia di contesto. Un PDF viene letto come immagine, costoso a ogni interrogazione. Il Markdown è testo: leggibile, grep-abile, riusabile a costo minimo.
- Guardrail anti-allucinazione. L'indice è il presupposto del protocollo quote-then-claim: non si cita ciò che non è nell'indice.
I due livelli dell'indice
- Strutturale (concetto → fonte : pagina): PARTE / SEZIONE / CAPITOLO con il loro titolo-concetto e la pagina di inizio nel corpo. Serve a navigare per argomento.
- Citazionale (Rv → fonte : pagina → massima): ogni sentenza realmente presente, con numero Rv, pagina e citazione testuale. È il backbone del grounding.
L'indice master vive in
KNOWLEDGE_BASE/02_GIURISPRUDENZA/_MD_INDICIZZATO/INDICE_GIURISPRUDENZA_PENALE.md.
I Markdown convertiti stanno nella stessa cartella.
Come si converte un documento
# 1) Converti un PDF (testo nativo) in MD + indice JSON
python3 skills/penalista-archivio/scripts/converti_indicizza.py \
<input.pdf> KNOWLEDGE_BASE/02_GIURISPRUDENZA/_MD_INDICIZZATO/ \
--titolo "Etichetta leggibile della fonte"
# 2) Rigenera l'indice master unendo tutti gli *.index.json presenti
python3 skills/penalista-archivio/scripts/genera_indice_master.py \
KNOWLEDGE_BASE/02_GIURISPRUDENZA/_MD_INDICIZZATO/ \
KNOWLEDGE_BASE/02_GIURISPRUDENZA/_MD_INDICIZZATO/INDICE_GIURISPRUDENZA_PENALE.md
Dipendenza: pdfplumber (pip install pdfplumber).
Limiti — leggere con attenzione
- NON fa OCR. Se il PDF è una scansione (poco testo estraibile) lo script si ferma e lo segnala: va prima passato per un OCR. Non convertire mai una scansione "alla cieca".
- L'estrazione strutturale è tarata sulle Rassegne del Massimario (PARTE/SEZIONE/CAPITOLO).
Su documenti con struttura diversa l'indice strutturale può essere parziale: in quel caso
la ricerca per concetto si fa con
grepdirettamente sul.md. - L'indice citazionale riconosce il formato Cassazione
Sez. X, n. … del gg/mm/aaaa, Nome, Rv. …. Citazioni con formattazione anomala possono sfuggire: vanno aggiunte a mano se rilevanti. - Tutto il prodotto è materiale di lavoro: la conversione non sostituisce la lettura della fonte ufficiale da parte del difensore.
What ships with it
5 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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.
- 12d ago First seen · 119 lines · 228 tokens per session scan A 3ac0694195fc
penalista-archivio is a skill published in the GitHub repository Synthos-Logic/penalista-italia (1 stars, last pushed 1mo ago), licensed MIT. It adds 228 tokens to every session and 1,933 once invoked, about $0.0011 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 skills, from other repositories
doc-reader
A document parser for Korean government, court, and business files, including HWP, HWPX, PDF, XLSX, and DOCX. It converts them to Markdown while extracting tables, form fields, metadata, and document differences.
legal-contract-review
A tool for reviewing and drafting legal documents such as contracts, terms of service, privacy policies, and service-level agreements (SLAs). An SLA is a contract that defines expected service performance and remedies.
hr-draft-offer
A skill for drafting offer letters and employment contracts. It helps state pay, benefits, working conditions, start dates, probation, insurance deductions, and stock-option terms.
official-doc
Produce a 31C official document (board resolution, formal notice, letter of position, certificate of authority) using the locked corporate template. Authoritative voice, declarative language, reference numbering, and official seal block. Renders to PDF + DOCX. Trigger when the user says "board resolution", "formal…
partnership-doc
Produce a 31C partnership document (MOU, Letter of Intent, or term sheet) using the locked corporate template. Legal-adjacent document defining mutual obligations, governance, territory, exclusivity, and confidentiality between 31 Concept and a counterparty. Renders to PDF + DOCX. Trigger when the user says "MOU"…
vlm-ocr
OCR scanned or image-only corpora with vision-language models, in three phases. Use evaluate to compare candidate OCR systems against stratified human ground truth and pick one on measured CER/WER, run to build the production pipeline (model selection, image handling, prompts, architecture, batching, accuracy…