interview-preplisted

# Interview Prep — Skill Instructions ## What this skill does Given a candidature (resolved by company name + role title, or a free-text reference), produce one Obsidian `entretien` note with `categorie: "Préparation"` in `CRM-JobSearch/Entretiens/`. The body is ALWAYS the same five H2 sections, in this exact order: 1. **Société** 2. **Résumé de la job description** 3. **Pitch** (profile-positioned for the candidature's `target_profile`) 4. **Cas du parcours pertinents** 5. **Questions probables de l'interlocuteur** The Pitch section is grounded in the matching `profiles/p<n>_*.md` narrative file — concrete proofs, not generic claims. All vault I/O goes through the `jobsearch-vault` skill; the profile narrative files (which live in the mounted private folder, NOT in the Obsidian vault and NOT in the plugin package) are read with the standard `Read` tool. ## Step 0 — Inputs you need from the user Collect (and confirm) the following before doing anything else: 1. **Candidature reference** (required) — free text like "the Relief CTO one" or "Anthropic Applied AI". The skill searches by `entreprise` + `poste` in `CRM-JobSearch/Opportunites/`. 2. **Interview date** — default: tomorrow. Format `YYYY-MM-DD` for frontmatter; `DD-MM-YYYY` for the filename suffix. 3. **`type_entretien`** — default `"RH"`. One of `RH` / `Technique` / `Manager` / `Final` (drives the question set in Step 3, section 5). 4. **`interlocuteurs`** — default `["TBD"]`. Free-text list of names. Do not pr