Agent skill · AI & Agents

assistant-juridique-fr

Assistant juridique expert en droit français. Recherches, consultations, rédaction d'actes, contre-argumentaires, analyses de contrats et de pièces, veille, vérification et harmonisation de références. Produit des documents Word (computer use) ou des réponses conversationnelles structurées. MCP : OpenLegi, Themia, LegalDataHunter. Déclencher pour toute question juridique, tout concept de droit français (contrat, responsabilité, jurisprudence, code civil, tribunal, indemnisation), toute question sur le droit de l'UE, la CJUE ou la CEDH, et toute demande de vérification ou harmonisation de référ

lawve-aigithub.com/lawve-aiGitHub ↗
claude-codeships scriptsNOASSERTION
Install
npx skills add lawve-ai/awesome-legal-skills --skill assistant-juridique-christophe-quezel-ambrunaz --agent claude-code

Same command for any agent — swap --agent for codex, cursor, copilot.

Facts
Files in the skill folder: 29
SKILL.md size: 39 KB
Bundled scripts: yes
Path: skills/assistant-juridique-christophe-quezel-ambrunaz/SKILL.md
Open the folder on GitHub →
Where it comes from
Stars: 618
Language: Python

Weekly change comes from our own snapshots, not the repository page — it measures attention, not adoption.

Review
written from the skill's own SKILL.md · Aug 5, 2026

What it does

Assists with French law across all branches, European and international law from a French perspective. It conducts research, provides consultations, drafts acts, builds counter-arguments, analyzes contracts and documents, performs vigilance, verifies and harmonizes bibliographic references, and produces Word documents or structured conversational responses. It prompts for any legal question or concept and triggers for UE/ECtHR/ECJ topics, including verification/harmonization of references.

How it works

The skill defines an execution environment detection at startup (COWORK, CHAT_CU, CHAT) with corresponding capabilities and persistence. It follows a five-step rule for sourcing references: Search → Find → Read content → Verify content supports the assertion → Cite, refusing to cite unverified references. It enforces formal tracing and requires verifiable identifiers for doctrinal references. It guides the user through a multi-step search process across normative texts, jurisprudence, and doctrine using specified tools and scripts (e.g., OpenLegi, Themia, LegalDataHunter, HAL, and dedicated scripts). It mandates a formal, iterative sourcing workflow, including a mandatory doctrinal search script and a citation verification loop. It includes a structured workflow for task routing (Tasks 1–8 production and Tasks 9–10 references) and a mandatory playbook check before most tasks. It also instructs a file scan step at task onset to classify available documents (pieces, templates, reference documents).

When to use it

Use whenever handling a legal question in French law (contracts, liability, jurisprudence, Civil Code, tribunals, indemnities) or any EU/ECtHR/ECJ query, including requests to verify or harmonize bibliographic references. The workflow explicitly triggers for legal questions and reference verification tasks and adapts based on environment (COWORK, CHAT_CU, CHAT).

What it can touch

The skill references tools and scripts such as OpenLegi, Themia, and LegalDataHunter, plus the scripts for doctrine search (scripts/doctrine_search.py, scripts/hal_search.py) and web_search for non-France sources. It prescribes using a specific workflow of commands and tool invocations (e.g., OpenLegi:rechercher_code, LegalDataHunter:search, web_search). It interacts with the filesystem in environments that provide a persistent filesystem (COWORK) or ephemeral storage (CHAT_CU) depending on mode.

Caveats

It enforces strict anti-hallucination rules: no citation without verified support, and every reference must be traceable with an official link (Légifrance or equivalent) and an identifier. If certain tools are unavailable, it switches to alternative sources (web_search) and informs the user of limitations. It may halt if core environmental prerequisites are unmet before starting a task.

From the SKILL.md

# Assistant Juridique FR ## §0 — Détection d'environnement **Au lancement, déterminer le mode d'exécution :** | Mode | Détection | Capacités | |---|---|---| | **COWORK** | Dossier de projet Cowork présent | Filesystem persistant, scripts, édition XML, Word, reprise inter-session | | **CHAT_CU** | Computer use activé, pas de dossier Cowork | Filesystem éphémère (`/mnt/user-data/uploads/`), scripts, édition XML, Word — mais pas de persistance entre conversations | | **CHAT** | Ni computer use ni Cowork | MCP (OpenLegi, Themia), web_search — pas de filesystem, pas de Word | **Règle de routage** : chaque fichier de tâche (`references/tache-*.md`) définit ses **pré-requis environnement** en en-tête. Si l'environnement courant ne satisfait pas les pré-requis, **interrompre avant de commencer** et demander à l'utilisateur d'activer computer use ou de basculer sur Cowork. Ne pas tenter d'exécuter en mode dégradé une tâche qui nécessite le filesystem. **Chemin des fichiers utilisateur** : - COWORK : dossier de travail du projet - CHAT_CU : `/mnt/user-data/uploads/` - CHAT : fichiers dans la fenêtre de contexte uniquement ## §1 — Identité et paradigme Assistant juridique expert couvrant l'en

What's inside
Steps it walks through
  1. §0 — Détection d'environnement
  2. §1 — Identité et paradigme
  3. §2 — Règle cardinale : anti-hallucination
  4. §3 — Séquence de recherche
  5. §4 — Scan des fichiers disponibles
  6. §5 — Routage des tâches
  7. Tâche 0 — Playbook juridique (cadrage préalable)
  8. Tâches 1-8 — Production documentaire
  9. Tâches 9-10 — Références (séparables, combinables)
  10. Données jurimétriques et Cour de cassation (Themia)
  11. Articulation avec les autres compétences (renvois)
  12. §6 — Format de sortie
  13. §7 — Application de la loi dans le temps
  14. §8 — Qualification et hiérarchie des normes
Ships with 24 files
  • README.md
  • evals/scenarios-non-regression.md
  • references/checklist-pre-livraison.md
  • references/format-citations.md
  • references/guide-hal.md
  • references/guide-legaldatahunter.md
  • references/guide-openlegi.md
  • references/guide-themia.md
  • references/principes-cardinaux.md
  • references/sources-fiables.md
  • references/tache-0-playbook.md
  • references/tache-1-recherches.md
  • references/tache-10-harmonisation-references.md
  • references/tache-2-cas-pratique.md
  • references/tache-3-redaction-acte.md
  • references/tache-4-analyse-pieces.md
  • references/tache-5-contre-argumentaire.md
  • references/tache-6-analyse-contrat.md
  • references/tache-7-veille-juridique.md
  • references/tache-8-analyse-article.md
  • references/tache-9-verification-references.md
  • scripts/doctrine_search.py
  • scripts/extract_references.py
  • scripts/format_citation.py
first 24 of 29
More from awesome-legal-skills
All skills →
About this skill
What does the assistant-juridique-fr skill do?

Assistant juridique expert en droit français. Recherches, consultations, rédaction d'actes, contre-argumentaires, analyses de contrats et de pièces, veille, vérification et harmonisation de références. Produit des documents Word (computer use) ou des réponses conversationnelles structurées. MCP : OpenLegi, Themia, LegalDataHunter. Déclencher pour toute question juridique, tout concept de droit français (contrat, responsabilité, jurisprudence, code civil, tribunal, indemnisation), toute question sur le droit de l'UE, la CJUE ou la CEDH, et toute demande de vérification ou harmonisation de référ

How do I install it?

Run `npx skills add lawve-ai/awesome-legal-skills --skill assistant-juridique-christophe-quezel-ambrunaz --agent claude-code` — it drops the skill into your project so the agent can pick it up. Swap the --agent value for codex, cursor or copilot if you use one of those.

Where does this skill come from?

From lawve-ai/awesome-legal-skills, a repository with 618 stars. We read it straight from the repository tree rather than a submitted listing, so what you see here is what is actually published.

Is a popular skill a good skill?

Not necessarily. Stars measure attention, not adoption — a repository can trend for a week and be abandoned. That is why we show the weekly change from our own snapshots next to the total, instead of a single flattering number.

Keep going