# Osf Datasets

> Access Open Science Framework (OSF) datasets via CLI using osfclient and the OSF v2 API. Load when downloading research datasets, searching public OSF projects, or uploading data to OSF repositories.

- Skill: `plurigrid/osf-datasets` (Agent Skill, multi-file: 4 files)
- Install (CLI): `npx skillmds@latest add plurigrid/osf-datasets`
- Raw SKILL.md: https://api.skillmd.com/api/skills/plurigrid/osf-datasets/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Research & Search
- Author: plurigrid (https://skillmd.com/u/plurigrid)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/plurigrid/osf-datasets

---


# OSF Datasets

## Setup

```bash
pip install osfclient
```

## Authentication

```bash
# Option A: Personal Access Token (preferred) -- generate at osf.io/settings/tokens
export OSF_TOKEN=<pat>

# Option B: credentials
export OSF_USERNAME=user@example.com
export OSF_PASSWORD=pass
```

Or per-directory `.osfcli.config`:

```ini
[osf]
username = user@example.com
project = abc12
```

## Project ID

The 5-character alphanumeric slug from any `osf.io/<id>` URL. Example: `osf.io/abc12` -> project ID is `abc12`.

## Core Commands

| Command | Purpose |
|---|---|
| `osf -p <id> ls` | List all files |
| `osf -p <id> clone [dir]` | Download entire project |
| `osf -p <id> fetch osfstorage/path.csv local.csv` | Download single file |
| `osf -p <id> upload local.csv osfstorage/path.csv` | Upload file |
| `osf -p <id> geturl osfstorage/path.csv` | Get web URL |

Remote paths are prefixed with the storage provider, typically `osfstorage/`.

## Direct API Access

Base URL: `https://api.osf.io/v2/`

Use `--globoff` with curl -- OSF query params use `[]` which curl interprets as glob ranges.

```bash
# List project files
curl -sL --globoff "https://api.osf.io/v2/nodes/<id>/files/osfstorage/" | jq '.data[].attributes.name'

# Get project metadata
curl -sL --globoff "https://api.osf.io/v2/nodes/<id>/" | jq '.data.attributes'

# Search public nodes by title
curl -sL --globoff "https://api.osf.io/v2/nodes/?filter[title]=keyword&page[size]=20" | jq '.data[] | {id, title: .attributes.title}'

# Browse subfolder (use folder ID from parent listing)
curl -sL --globoff "https://api.osf.io/v2/nodes/<id>/files/osfstorage/<folder_id>/" | jq '.data[].attributes.name'

# With auth
curl -sL --globoff -H "Authorization: Bearer $OSF_TOKEN" "https://api.osf.io/v2/nodes/<id>/files/osfstorage/"
```

File download links are in `data[].links.download` -- follow redirects with `curl -L -o`.

## Helper Scripts

- `scripts/osf-search.sh <keyword> [page_size]` -- search public OSF projects by keyword
- `scripts/osf-browse.sh <project_id> [path]` -- browse project files via API (no osfclient needed)

## Guidelines

- Public projects need no auth; private ones require `OSF_TOKEN` or credentials
- `clone` mirrors the remote directory structure locally
- `fetch` requires both remote and local path arguments
- The API paginates at 10 items by default; use `?page[size]=100` for larger listings
- Rate limits apply -- add brief sleeps for batch operations