agentleFS
GeekFamilyCorpGeekFamilyCorp/TricorderKit/skills/docmancer/SKILL.md

docmancer

Skill32 starsChanged 15 days agoMITView on GitHub

Version 0.1.0 — 2026-05-18 Génération automatique de notes Obsidian et documents Markdown structurés. S'appuie sur obsidian-agent-layer pour le CRUD vault et les templates.

1 other project uses a version of this.

Adopted by

Most recently taken up by

The file

Files
SKILL.md

Skill: docmancer — Document & Note Generator

Version 0.1.0 — 2026-05-18 Génération automatique de notes Obsidian et documents Markdown structurés. S'appuie sur obsidian-agent-layer pour le CRUD vault et les templates.


Triggers

/tk:docmancer
/tk:doc
docmancer
génère une note
crée une fiche
indexe dans obsidian
note obsidian pour
fiche projet
fiche article
fiche personne
sauvegarde dans le vault
persiste dans obsidian

Prérequis

  1. plugins/obsidian-agent-layer/ présent et opérationnel
  2. MCP obsidian-claude-vault connecté (vault TricorderKit)
  3. MCP obsidian-my-domain-project connecté si une note cible le vault d'un linked_project

Instructions agent

Séquence d'exécution

ÉTAPE 1 — Intent extraction
  → Extraire : type (project|paper|person|organization|concept|source|generic), content, vault_target, dry_run
  → vault_target par défaut : "claude-vault"
  → Pour les entités d'un domaine métier : vault_target = "my-domain-project" (linked_project)

ÉTAPE 2 — Sélection template (via note_builder.py)
  | Type       | Template               | Vault cible        |
  |------------|------------------------|-------------------|
  | project      | templates/project.md      | claude-vault ou linked_project |
  | paper        | templates/paper.md        | claude-vault ou linked_project |
  | person       | templates/person.md       | claude-vault ou linked_project |
  | organization | templates/organization.md | claude-vault ou linked_project |
  | concept      | templates/concept.md      | claude-vault       |
  | source       | templates/source.md       | claude-vault       |
  | session    | templates/session.md   | claude-vault       |
  | generic    | templates/generic.md   | claude-vault       |
  | decision   | templates/decision.md  | claude-vault       |

ÉTAPE 3 — Vérification existence (anti-doublon R7)
  python plugins/obsidian-agent-layer/vault_router.py \
    --action check --path "<note_path>" --vault <vault>
  → Si note existe : proposer mise à jour, non réécriture

ÉTAPE 4 — Construction frontmatter + body
  python plugins/obsidian-agent-layer/note_builder.py \
    --type <type> --data '<json_content>' [--dry-run]

ÉTAPE 5 — Écriture vault (MCP ou obsidian_client.py)
  python plugins/obsidian-agent-layer/obsidian_client.py \
    --action write --path "<note_path>" --content '<content>' --vault <vault>

ÉTAPE 6 — Output contractuel JSON

Templates disponibles

Template project (frontmatter)

---
type: project
title: "<nom du projet>"
url: "<dépôt ou site officiel>"
maintainer: "<organisation ou personne>"
license: "<licence>"
language: "<langage principal>"
status: active|maintenance|archived
latest_release: "<version>"
topics: []
tags: ["#project"]
reliability: confirmed|probable|to_verify|incomplete
sources: []
created: "<YYYY-MM-DD>"
updated: "<YYYY-MM-DD>"
---

Template session (frontmatter)

---
type: session-log
date: "<YYYY-MM-DD>"
project: tricorderkit
version: "<version>"
tags: ["#session", "#tricorderkit"]
status: completed
---

Règles importantes

  • Anti-doublon (R7) : toujours vérifier existence avant création
  • Frontmatter obligatoire : toute note doit avoir un frontmatter YAML valide
  • Atomicité : 1 note = 1 entité (pas de notes multi-sujets)
  • Taille cible : 100–500 tokens par note (principe CLAUDE.md)
  • Fiabilité : toujours indiquer le niveau (confirmed|probable|to_verify|incomplete)
  • Dry-run (R3) : activer --dry-run pour prévisualiser avant écriture réelle
  • Décisions : loguer dans .planning/DECISIONS.md si la note implique un choix archi

Commandes de référence

# Construction note projet (dry-run)
python plugins/obsidian-agent-layer/note_builder.py \
  --type project \
  --data '{"title":"Qdrant","url":"https://github.com/qdrant/qdrant","license":"Apache-2.0"}' \
  --dry-run

# Vérification existence
python plugins/obsidian-agent-layer/vault_router.py \
  --action check --path "Projects/Qdrant.md" --vault claude-vault

# Écriture réelle
python plugins/obsidian-agent-layer/obsidian_client.py \
  --action write --path "Projects/Qdrant.md" --vault claude-vault \
  --content '<frontmatter+body>'

# Mise à jour note existante
python plugins/obsidian-agent-layer/obsidian_client.py \
  --action patch --path "Projects/Qdrant.md" \
  --field "latest_release" --value "1.12.0"

Format de sortie contractuel

{
  "status": "success",
  "skill_name": "docmancer",
  "skill_version": "0.1.0",
  "timestamp": "2026-05-18T10:00:00Z",
  "output": {
    "action": "created",
    "note_path": "Projects/Qdrant.md",
    "vault": "claude-vault",
    "type": "project",
    "template_used": "templates/project.md",
    "frontmatter_keys": ["title", "url", "license", "status", "reliability"],
    "reliability": "confirmed",
    "dry_run": false,
    "next_steps": ["Lier à la fiche organisation", "Ajouter la dernière version publiée"]
  },
  "metadata": {
    "tokens_estimated": 350,
    "vault_path": "${OBSIDIAN_VAULT_PATH}/Projects/Qdrant.md"
  }
}

Gestion d'erreurs

ErreurCauseAction
Note already existsDoublon détectéProposer patch, non réécriture
obsidian-agent-layer introuvablePlugin non installéVérifier plugins/obsidian-agent-layer/
MCP obsidian-my-domain-project timeoutVault non connectéFallback claude-vault + warning
Frontmatter invalideDonnées manquantesDemander les champs requis à l'utilisateur
Template inconnuType non supportéUtiliser template generic

Pipeline rtk → docmancer

Ce skill est conçu pour être appelé directement après rtk :

/rtk "Qdrant"
  → résultats scorés
  → /tk:docmancer --type project --data <rtk_output>
    → note créée dans le vault cible

Output caveman pour pipeline (R15) :

{
  "status": "ok",
  "task": "docmancer_write",
  "data": {"note": "Projects/Qdrant.md", "vault": "claude-vault", "action": "created"},
  "tokens_used": 350,
  "next_action": "review_queue_add"
}

TricorderKit v0.8 — GeekFamilyCorp — 2026-05-18

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.