# /setup Path A Documents Folder Structure for Importing Materials

> Understand the /setup Path A documents folder structure for importing materials into your AI job search assistant. Learn how the documents directory populates candidate profiles and skills.

- Repository: [Mads Lorentzen/ai-job-search](https://github.com/MadsLorentzen/ai-job-search)
- Tags: how-to-guide
- Published: 2026-07-08

---

**TLDR:** **/setup Path A** is the documents-folder workflow that scans the `documents/` directory to automatically populate your candidate profile and skill files for the AI job search assistant.

The `MadsLorentzen/ai-job-search` repository provides a structured workflow for managing job applications using Claude Code. When you run the `/setup` command and select **Path A**, Claude reads your existing career documents from a specific folder hierarchy defined in [`documents/README.md`](https://github.com/MadsLorentzen/ai-job-search/blob/main/documents/README.md) to build a comprehensive candidate profile.

## What Is /setup Path A?

Path A is the recommended onboarding workflow when you have multiple source documents—such as a master CV, LinkedIn export, diploma PDFs, and reference letters—that you want Claude to ingest in a single operation. Unlike Path B (single CV upload) or Path C (interactive interview), Path A expects a predefined folder structure under the top-level `documents/` directory.

According to the repository's [`SETUP.md`](https://github.com/MadsLorentzen/ai-job-search/blob/main/SETUP.md) guide, this path is the most repeatable: you can add new files at any time and re-run `/setup`, and Claude will only update the sections that have changed.

## The Documents Folder Hierarchy

The folder structure is fixed and documented in [`documents/README.md`](https://github.com/MadsLorentzen/ai-job-search/blob/main/documents/README.md). Claude expects the following hierarchy:

```

documents/
├── cv/                # Your master CV (PDF or .tex)

├── linkedin/          # LinkedIn profile export (PDF)

├── diplomas/          # Degree certificates / transcripts (PDF)

├── references/        # Reference letters (PDF, .md, .txt)

├── applications/      # Past applications, one sub-folder per role

│   └── <company>_<role>/
│       ├── job_posting.md   # Original posting text

│       ├── cover_letter.tex # The cover letter you sent

│       ├── cv_draft.tex     # The CV variant you sent

│       └── outcome.md       # Result + notes

└── README.md          # Documentation for this structure

```

Each subdirectory serves a specific purpose in the profile generation process. The `applications/` folder uses a nested structure where each subfolder represents a specific role, allowing Claude to analyze your application history and outcomes.

## How Path A Processes Your Materials

When you invoke `/setup` and select Path A, the command defined in [`.claude/commands/setup.md`](https://github.com/MadsLorentzen/ai-job-search/blob/main/.claude/commands/setup.md) executes a four-stage pipeline:

1. **Scanning**. Claude recursively scans each sub-folder (`cv/`, `linkedin/`, `diplomas/`, `references/`, `applications/`) to identify supported file types.

2. **Extraction**. Built-in parsers process PDF, LaTeX, and Markdown files to extract structured data including experience, education, skills, contact information, and reference quotes.

3. **Merging**. The extracted data merges with any existing profile files. Claude prompts only when conflicts arise between new and existing information.

4. **Population**. The system populates the canonical profile files—including [`CLAUDE.md`](https://github.com/MadsLorentzen/ai-job-search/blob/main/CLAUDE.md), [`01-candidate-profile.md`](https://github.com/MadsLorentzen/ai-job-search/blob/main/01-candidate-profile.md), and [`02-behavioral-profile.md`](https://github.com/MadsLorentzen/ai-job-search/blob/main/02-behavioral-profile.md)—that downstream commands like `/apply` and `/interview` rely on.

## Setting Up the Folder Structure

Create the required hierarchy using command-line tools before running `/setup`.

**Create directories (Bash/Zsh):**

```bash
mkdir -p documents/{cv,linkedin,diplomas,references,applications}
touch documents/README.md

```

**Create directories (PowerShell):**

```powershell
New-Item -ItemType Directory -Path documents\cv,documents\linkedin,documents\diplomas,documents\references,documents\applications
New-Item -ItemType File -Path documents\README.md

```

**Populate with your materials:**

```bash

# Add master CV

cp ~/Downloads/my_cv.pdf documents/cv/

# or for LaTeX source

cp ~/Downloads/my_cv.tex documents/cv/

# Add LinkedIn export

cp ~/Downloads/LinkedIn_Profile.pdf documents/linkedin/

# Add a past application

mkdir -p documents/applications/acme_ml_engineer
cat > documents/applications/acme_ml_engineer/job_posting.md <<'EOF'
[Paste the full job description here]
EOF
cp ~/Downloads/acme_cover.tex documents/applications/acme_ml_engineer/cover_letter.tex
cp ~/Downloads/acme_cv.tex documents/applications/acme_ml_engineer/cv_draft.tex

```

**Run the setup:**

```bash
claude                # Start Claude Code REPL

/setup                # Invoke the onboarding command

# Select Path A when prompted

```

Claude will summarize the extracted data and ask you to confirm any conflicts before finalizing the profile.

## Key Configuration Files

Understanding these files helps you customize the workflow:

- **[`documents/README.md`](https://github.com/MadsLorentzen/ai-job-search/blob/main/documents/README.md)**: The definitive specification of the documents folder hierarchy and supported file types.
- **[`SETUP.md`](https://github.com/MadsLorentzen/ai-job-search/blob/main/SETUP.md)**: The high-level onboarding guide that explains the three `/setup` paths and command flow.
- **[`.claude/commands/setup.md`](https://github.com/MadsLorentzen/ai-job-search/blob/main/.claude/commands/setup.md)**: The actual Claude Code command definition that implements the interactive Path A logic.
- **[`CLAUDE.md`](https://github.com/MadsLorentzen/ai-job-search/blob/main/CLAUDE.md)**: The consolidated candidate profile generated by `/setup` and used by other commands.
- **`cv/main_example.tex`**: Example LaTeX CV template that demonstrates the expected format for CV files.

## Summary

- **/setup Path A** requires a specific `documents/` folder structure defined in [`documents/README.md`](https://github.com/MadsLorentzen/ai-job-search/blob/main/documents/README.md).
- The hierarchy includes dedicated folders for your CV, LinkedIn profile, diplomas, references, and past applications.
- Claude extracts structured data from PDF, LaTeX, and Markdown files to populate [`CLAUDE.md`](https://github.com/MadsLorentzen/ai-job-search/blob/main/CLAUDE.md) and skill profiles.
- This path is idempotent—re-running `/setup` only updates changed sections, making it ideal for evolving career documents.

## Frequently Asked Questions

### What file formats does Path A support?

Path A supports **PDF**, **LaTeX (.tex)**, and **Markdown (.md)** files, with plain text (.txt) accepted for reference letters. The parsers extract text and structure from these formats to populate your profile. Ensure scanned PDFs are text-searchable for optimal extraction.

### Can I re-run /setup Path A after adding new documents?

Yes. The Path A workflow is designed to be **repeatable and idempotent**. When you re-run `/setup` and select Path A, Claude scans the `documents/` folder again and merges new information with your existing profile in [`CLAUDE.md`](https://github.com/MadsLorentzen/ai-job-search/blob/main/CLAUDE.md), prompting only when conflicts arise between old and new data.

### What happens if I don't have all the subfolders?

You only need to create the subfolders relevant to your materials. If a directory like `diplomas/` or `references/` is empty or missing, Claude skips it and proceeds with the available data. However, maintaining the full structure ensures comprehensive profile generation.

### Where does Claude store the extracted profile data?

Claude populates several files in the `.claude/skills/job-application-assistant/` directory, including [`CLAUDE.md`](https://github.com/MadsLorentzen/ai-job-search/blob/main/CLAUDE.md), [`01-candidate-profile.md`](https://github.com/MadsLorentzen/ai-job-search/blob/main/01-candidate-profile.md), and [`02-behavioral-profile.md`](https://github.com/MadsLorentzen/ai-job-search/blob/main/02-behavioral-profile.md). These files serve as the single source of truth for downstream commands like `/apply` and `/interview`.