CVE-2026-47277: Runtipi Arbitrary File Read via Symlink in App Logo Endpoint

Technical summary

The vulnerability stems from insufficient path traversal protection in the unauthenticated logo-serving endpoint. Runtipi performs a lexical (string-based) check of file paths before reading, but this validation occurs before the operating system resolves symbolic links. An attacker controlling a Git app-store repository can place a symlink at the expected logo location (e.g., metadata/logo.jpg) pointing to sensitive files. When a user or automated process requests the logo, Runtipi reads the symlink target without further validation, exposing files outside the intended repository directory. The flawed logic fails to account for post-validation symlink resolution, a common path-traversal mistake. Affected paths may include /data/.env, /data/state/seed, and application configuration files stored within the Runtipi container.

Business impact

This vulnerability poses a direct confidentiality risk to Runtipi deployments. Exposed .env files typically contain database credentials, API keys, and JWT signing secrets that could enable lateral movement or authentication bypass in broader infrastructure. Seed files and cryptographic material are especially valuable to attackers seeking persistent access or certificate forgery. For personal homeserver operators, exposure of operational logs could reveal system architecture, running services, or user activity patterns. If Runtipi is used in shared environments or integrated with organizational infrastructure, the credential leakage risk escalates significantly. The unauthenticated nature of the endpoint means no attacker capability assumptions are required beyond basic network access.

Affected systems

Runtipi versions 4.9.1, 4.9.2, and 4.9.3 are confirmed vulnerable. The vulnerability requires the targeted Runtipi instance to have cloned a malicious app-store repository and attempted to load logos from it; however, because Runtipi's marketplace functionality is a core feature, most active deployments are likely at risk if they interact with any third-party app stores. Version 4.10.0 and later contain the fix.

Exploitability

Exploitability is straightforward. An attacker needs only to create a public Git repository formatted as a Runtipi app store and include a symlink at the logo location pointing to /data/.env or another sensitive target. Once a Runtipi instance adds this malicious app store and attempts to serve the logo (whether manually or through automatic app discovery), the unauthenticated endpoint leaks the target file content over HTTP. No authentication, user interaction during the attack, or special capabilities are required to exploit, though the CVSS rating of 6.5 (Medium) reflects that user interaction is factored into the scoring—specifically, a user must trigger the logo load by viewing the app or Runtipi discovering it. The technical bar for creating a malicious app store is low.

Remediation

Upgrade to Runtipi version 4.10.0 or later, which implements proper symlink resolution checks in the logo-serving endpoint. Before patching, if you cannot immediately upgrade, consider restricting network access to your Runtipi instance and avoiding untrusted app-store repositories. Do not add external or third-party app stores from unknown sources until you have confirmed your version is patched. Additionally, rotate any secrets (JWT keys, database credentials, API tokens) that may have been exposed through logs or .env files in unpatched instances.

Patch guidance

Apply the official Runtipi update to version 4.10.0 or later. Verify the release notes and checksums from the official Runtipi repository to ensure authenticity. If you are running a self-hosted or containerized deployment, rebuild your container image with the patched version and restart services. For managed Runtipi instances, check with your provider for automatic patching status. After patching, review application logs for any suspicious requests to the logo endpoint (typically /logo endpoints or similar paths) from unexpected sources, which may indicate prior exploitation attempts.

Detection guidance

Monitor HTTP access logs for repeated requests to logo-serving endpoints (typically paths like /api/app/*/logo or /apps/*/logo) from unusual IP addresses or user agents, especially requests that precede the patching date. Watch for 200 responses returning unusually large payloads (symlink targets often pull in larger system files than expected small image data). Review Runtipi application logs for warnings about symlinks or path validation failures. If you have added third-party app stores, audit their repository contents for suspicious symlinks using `find repo -type l` or equivalent commands. Examine the /data/.env and /data/state/seed files on affected instances for signs of compromise (unexpected access timestamps, missing or modified content).

Why prioritize this

Although rated CVSS 6.5 (Medium), this vulnerability merits prompt attention due to its unauthenticated, easily exploitable nature and the high sensitivity of exposed artifacts (JWT secrets, database credentials). The attack surface—any app store a user adds—is broad and relatively invisible. Credential theft from .env files can lead to full container compromise, making this a stepping stone to further attacks. Organizations running Runtipi for critical personal infrastructure or in connected networks should prioritize patching within days rather than weeks.

Risk score, explained

The CVSS 6.5 score reflects a high-impact confidentiality breach (C:H) with no authentication required (PR:N) and low attack complexity (AC:L), balanced against the requirement for user interaction (UI:R)—specifically, a user must trigger the logo load via the marketplace. The Remote (AV:N) vector confirms network exploitability. No integrity or availability impact is scored. In real-world context, the score may underrepresent risk if your instance uses public or semi-trusted app stores, as the attack surface is effectively larger than a typical authenticated file-read.

Frequently asked questions