GHSA-239w-m3h6-ch8v CVE-2026-54094 GO-2026-5055

File Browser: Symlink following lets scoped users read, overwrite, and share files outside their filebrowser scope

High Go

Severity: 6.8 / 10; High · CVSS 3.1
Exploit chance (EPSS): 0.50%; Chance of exploitation in the next 30 days. Higher than 43.03% of scored vulnerabilities.
Known exploited: No; Not in CISA's catalog
Weakness: CWE-22 Improper Limitation of a Pathname to a Restricted Directory ('Path Traversal'); CWE-59 Improper Link Resolution Before File Access ('Link Following')
Fix: 2.63.14; Upgrade github.com/filebrowser/filebrowser/v2 to this version or later

Description

Summary

File Browser enforces per-user scope with afero.NewBasePathFs(afero.NewOsFs(), scope), set up in users/users.go. This blocks lexical ../ traversal, but it does not stop the HTTP file handlers from following symbolic links before they open, serve, write, share, or list a file. As a result, a scoped user — and in some cases an unauthenticated public-share recipient — can cross the intended scope boundary by following a symlink whose path is lexically inside their scope but whose target is outside it.

Two distinct shapes are covered here:

Variant 1 — symlink as the final path component. A symlink that lives inside the user's scoped tree and points to a file under the server root but outside the scope. The handlers record the symlink (IsSymlink) but then resolve and operate on the target anyway.
Variant 2 — file or directory reached through a symlinked ancestor. A regular file requested through a symlinked directory.

Read, write (including TUS resumable uploads), share creation, and public-share serving are all affected.

Impact

In a multi-user deployment, if a symlink (a file symlink for Variant 1, or a directory symlink for Variant 2) exists inside a restricted user's scoped tree and resolves to a location outside that scope but reachable by the server process, the boundary can be crossed. Concretely, a user holding only normal File Browser permissions can:

Read out-of-scope file contents and metadata via GET /api/raw/{path} and GET /api/resources/{path}.
Overwrite an out-of-scope target via POST /api/resources/{path}?override=true.
Overwrite or create an out-of-scope target via the TUS resumable upload path: POST /api/tus/{path}?override=true followed by PATCH /api/tus/{path}.
Create a public share for an out-of-scope target via POST /api/share/{path}, exposing it through GET /api/public/dl/{hash}.

For Variant 2, the same exposure reaches public-share recipients: a normal public directory share whose subtree contains a linked descendant lets an unauthenticated recipient read regular files behind the link, pull them into the share's archive download, and see the resolved target in directory listings.

This breaks the confidentiality and integrity guarantees that per-user scopes and password/anonymous shares are relied upon to provide, for any data the server process can reach.

Technical details

Users are rooted with afero.NewBasePathFs(afero.NewOsFs(), scope). Base-path rooting blocks lexical ../ traversal but does not prevent ordinary filesystem operations from following a symlink whose path is lexically inside the base.

The metadata layer records symlinks but does not consistently re-check the resolved target against the user's real scope:

In files/file.go, stat() calls LstatIfPossible, sets IsSymlink, and only invokes the WithinScope containment check when file.IsSymlink == true. For Variant 1, this guard (where present) covers the final-element symlink; on the commit tested for Variant 1 the handler still resolved the target with opts.Fs.Stat(opts.Path) and served it. For Variant 2, LstatIfPossible follows a symlinked ancestor and returns the leaf as a regular file (IsSymlink == false), so stat() returns early and the scope check never runs at all.
readListing in files/file.go follows symlink entries to display the target's metadata.
http/raw.go builds a file object for the requested path and serves non-directories; its archive walker getFiles follows symlinks via Stat/Open, pulling linked descendants into archive downloads.
http/resource.go writes request bodies with writeFile(d.user.Fs, r.URL.Path, ...), and the destination open follows symlinks.
http/tus_handlers.go (tusPostHandler, tusPatchHandler) calls MkdirAll/OpenFile on the request path directly with no containment check. Because a brand-new leaf does not stat an existing file, it skips the scope check entirely.
http/share.go stores a share for r.URL.Path without checking that the path is not a symlink escape; http/public.go later serves it for unauthenticated downloads (routed at http/http.go:90-91).
http/data.go applies dotfile and rule checks to the request path string, but never compares the resolved symlink target against the user's real scope.

Proof of concept

Variant 1 — symlink as final path component

Harness layout: server root is a temp directory; restricted user restricted is scoped to /u1 with create, modify, rename, share, and download permissions; a second scope /u2 holds the outside target /u2/secret.txt containing other-secret; and /u1/link-out is a symlink to /u2/secret.txt.

Confirmed bypasses (route-level tests against the real HTTP handlers):

GET /api/raw/link-out → 200 OK, body contains other-secret from /u2/secret.txt.
POST /api/resources/link-out?override=true → 200 OK, /u2/secret.txt changed to pwn.
POST /api/tus/link-out?override=true → 201, then PATCH /api/tus/link-out → 204, /u2/secret.txt changed.
POST /api/share/link-out → 200 OK, created a public share whose GET /api/public/dl/{hash} returned a body containing other-secret.

Minimal core of the read proof:

root := t.TempDir()
os.MkdirAll(filepath.Join(root, "u1"), 0755)
os.MkdirAll(filepath.Join(root, "u2"), 0755)
os.WriteFile(filepath.Join(root, "u2", "secret.txt"), []byte("other-secret"), 0644)
os.Symlink(filepath.Join(root, "u2", "secret.txt"), filepath.Join(root, "u1", "link-out"))

// restricted is a File Browser user scoped to /u1 with Download permission.
rr := authenticatedRequest(t, restricted, http.MethodGet, "/api/raw/link-out", nil)
if rr.Code != http.StatusOK || !strings.Contains(rr.Body.String(), "other-secret") {
    t.Fatalf("raw symlink exposed outside target: status=%d body=%q", rr.Code, rr.Body.String())
}

Variant 2 — file reached through a symlinked ancestor

Authenticated scoped user whose scope contains a directory symlink escape_link -> /srv/users/otheruser:

# The symlink itself is correctly blocked
GET /api/resources/escape_link              -> 403 Forbidden

# A regular file THROUGH the symlinked directory is not
GET /api/resources/escape_link/private.txt  -> 200 OK  {"content":"OTHER_USER_SECRET_DATA=...",...}
GET /api/raw/escape_link/private.txt        -> 200 OK  OTHER_USER_SECRET_DATA=...

# Create/overwrite THROUGH the symlinked directory (TUS)
POST  /api/tus/escape_link/injected.txt  (Upload-Length: 20) -> 201 Created
PATCH /api/tus/escape_link/injected.txt  (Upload-Offset: 0)  -> 204 No Content  (written into /srv/users/otheruser/)

Public directory share for /shared, where /shared/link -> ../private and private/secret.txt lives outside the share:

GET /api/public/dl/<hash>/link/secret.txt     -> 200 OK  symlink-secret
GET /api/public/share/<hash>/link/secret.txt  -> 200 OK  {"path":"/link/secret.txt", ...}

Requesting the whole share as an archive pulls link/secret.txt into the zip, and listing the share root exposes the link entry with its resolved target metadata.

Controls that held

The same harness confirmed that ordinary traversal is still rejected, so this is not generic ../ traversal:

GET /api/resources/../u2/secret.txt?checksum=sha256 did not succeed as the restricted user.
GET /api/resources/%2e%2e/u2/secret.txt did not succeed (encoded dot-dot).
POST /api/resources/../u2/new.txt did not create /u2/new.txt.
PATCH /api/resources/own.txt?action=rename&destination=/../u2/moved.txt did not move a file outside scope.

Affected code

users/users.go (scope setup); files/file.go (stat, readListing); http/raw.go (getFiles); http/resource.go (writeFile destination); http/tus_handlers.go (tusPostHandler, tusPatchHandler); http/share.go; http/public.go; http/http.go:90-91 (public routes); http/data.go (string-only path checks).

Remediation

Resolve symlinks and verify that the resolved target remains inside the user's real scoped root before any file operation — serving, sharing, writing, truncating, renaming, copying, or deleting. Specifically:

Call WithinScope (which resolves every path component with filepath.EvalSymlinks) for all paths in stat(), not only when the final element is a symlink. This closes the ancestor-symlink gap (Variant 2).
Add a WithinScope check before MkdirAll/OpenFile in tusPostHandler and tusPatchHandler, so a not-yet-existing leaf cannot skip containment.
Omit entries whose resolved target escapes the scope from readListing, and skip them in getFiles before stat/open/recursion.
Apply the same resolved-path check consistently to public share creation and public share serving.
As an alternative or defense-in-depth, reject symlinks for file operations unless an explicit administrator option enables them.

Add regression tests covering symlink reads, overwrites, TUS create/write, public shares (download, share-info, listing, and archive read paths), and the existing dot-dot controls — plus a positive test confirming that legitimately in-scope symlinks still resolve.

Limitations and non-claims

This is not generic ../ path traversal; dot-dot and encoded dot-dot controls held in the route-level tests.
This is not a proxy-auth confusion issue; the proofs use normal authenticated requests for a restricted user (and, for Variant 2's share case, an ordinary public-share recipient).
The proofs assume the relevant symlink already exists inside the scoped tree, or that another allowed workflow in the deployment can place it there — for example an SMB/NFS export, a Docker bind-mount, or an admin-created link. Web-UI-only creation of the symlink from scratch was not demonstrated.

Affected versions

Go

Package	Vulnerable	Fixed in
github.com/filebrowser/filebrowser	`<= 1.11.0`	No fix yet
github.com/filebrowser/filebrowser	`all versions`	No fix yet
github.com/filebrowser/filebrowser/v2	`< 2.63.14`	`2.63.14`

How to fix

Move to the fixed version or later with your package manager, update the lockfile, and run your tests. If the package only arrives through another dependency, upgrade that dependency first. If none of its releases pulls a fixed version yet, force the version with your package manager's override.

github.com/filebrowser/filebrowser/v2 Go → `2.63.14`

go

go get github.com/filebrowser/filebrowser/v2@v2.63.14 && go mod tidy

Severity breakdown

CVSS 3.1 6.8

CVSS:3.1/AV:N/AC:H/PR:L/UI:N/S:U/C:H/I:H/A:N

Attack vector: Network
Attack complexity: High
Privileges required: Low
User interaction: None
Scope: Unchanged
Confidentiality: High
Integrity: High
Availability: None

Weakness

CWE-22: Improper Limitation of a Pathname to a Restricted Directory ('Path Traversal')

The product uses external input to construct a pathname that is intended to identify a file or directory that is located underneath a restricted parent directory, but the product does not properly neutralize special elements within the pathname that can cause the pathname to resolve to a location that is outside of the restricted directory.

How to prevent it

Implementation: Assume all input is malicious. Use an "accept known good" input validation strategy, i.e., use a list of acceptable inputs that strictly conform to specifications. Reject any input that does not strictly conform to specifications, or transform it into something that does. When performing input validation, consider all potentially relevant properties, including length, type of input, the full range of acceptable values, missing or extra inputs, syntax, consistency across related fields, and conformance to business rules. As an example of business rule logic, "boat" may be syntactically valid because it only contains alphanumeric characters, but it is not valid if the input is only expected to contain colors such as "red" or "blue." Do not rely exclusively on looking for malicious or malformed inputs. This is likely to miss at least one undesirable input, especially if the code's environment changes. This can give attackers enough room to bypass the intended validation. However, denylists can be useful for detecting potential attacks or determining which inputs are so malformed that they should be rejected outright. When validating filenames, use stringent allowlists that limit the character set to be used. If feasible, only allow a single "." character in the filename to avoid weaknesses such as CWE-23, and exclude directory separators such as "/" to avoid CWE-36. Use a list of allowable file extensions, which will help to avoid CWE-434. Do not rely exclusively on a filtering mechanism that removes potentially dangerous characters. This is equivalent to a denylist, which may be incomplete (CWE-184). For example, filtering "/" is insufficient protection if the filesystem also supports the use of "\" as a directory separator. Another possible error could occur when the filtering is applied in a way that still produces dangerous data (CWE-182). For example, if "../" sequences are removed from the ".../...//" string in a sequential fashion, two instances of "../" would b
Architecture and Design: For any security checks that are performed on the client side, ensure that these checks are duplicated on the server side, in order to avoid CWE-602. Attackers can bypass the client-side checks by modifying values after the checks have been performed, or by changing the client to remove the client-side checks entirely. Then, these modified values would be submitted to the server.
Implementation: Inputs should be decoded and canonicalized to the application's current internal representation before being validated (CWE-180). Make sure that the application does not decode the same input twice (CWE-174). Such errors could be used to bypass allowlist validation schemes by introducing dangerous inputs after they have been checked. Use a built-in path canonicalization function (such as realpath() in C) that produces the canonical version of the pathname, which effectively removes ".." sequences and symbolic links (CWE-23, CWE-59). This includes: realpath() in C getCanonicalPath() in Java GetFullPath() in ASP.NET realpath() or abs_path() in Perl realpath() in PHP
Architecture and Design: Use a vetted library or framework that does not allow this weakness to occur or provides constructs that make this weakness easier to avoid [REF-1482].
Operation: Use an application firewall that can detect attacks against this weakness. It can be beneficial in cases in which the code cannot be fixed (because it is controlled by a third party), as an emergency prevention measure while more comprehensive software assurance measures are applied, or to provide defense in depth [REF-1481].
Architecture and Design, Operation: Run your code using the lowest privileges that are required to accomplish the necessary tasks [REF-76]. If possible, create isolated accounts with limited privileges that are only used for a single task. That way, a successful attack will not immediately give the attacker access to the rest of the software or its environment. For example, database applications rarely need to run as the database administrator, especially in day-to-day operations.
Architecture and Design: When the set of acceptable objects, such as filenames or URLs, is limited or known, create a mapping from a set of fixed input values (such as numeric IDs) to the actual filenames or URLs, and reject all other inputs. For example, ID 1 could map to "inbox.txt" and ID 2 could map to "profile.txt". Features such as the ESAPI AccessReferenceMap [REF-185] provide this capability.
Architecture and Design, Operation: Run the code in a "jail" or similar sandbox environment that enforces strict boundaries between the process and the operating system. This may effectively restrict which files can be accessed in a particular directory or which commands can be executed by the software. OS-level examples include the Unix chroot jail, AppArmor, and SELinux. In general, managed code may provide some protection. For example, java.io.FilePermission in the Java SecurityManager allows the software to specify restrictions on file operations. This may not be a feasible solution, and it only limits the impact to the operating system; the rest of the application may still be subject to compromise. Be careful to avoid CWE-243 and other weaknesses related to jails.
Architecture and Design, Operation: Store library, include, and utility files outside of the web document root, if possible. Otherwise, store them in a separate directory and use the web server's access control capabilities to prevent attackers from directly requesting them. One common practice is to define a fixed constant in each calling program, then check for the existence of the constant in the library/include file; if the constant does not exist, then the file was directly requested, and it can exit immediately. This significantly reduces the chance of an attacker being able to bypass any protection mechanisms that are in the base program but not in the include files. It will also reduce the attack surface.
Implementation: Ensure that error messages only contain minimal details that are useful to the intended audience and no one else. The messages need to strike the balance between being too cryptic (which can confuse users) or being too detailed (which may reveal more than intended). The messages should not reveal the methods that were used to determine the error. Attackers can use detailed information to refine or optimize their original attack, thereby increasing their chances of success. If errors must be captured in some detail, record them in log messages, but consider what could occur if the log messages can be viewed by attackers. Highly sensitive information such as passwords should never be saved to log files. Avoid inconsistent messaging that might accidentally tip off an attacker about internal state, such as whether a user account exists or not. In the context of path traversal, error messages which disclose path information can help attackers craft the appropriate attack strings to move through the file system hierarchy.
Operation, Implementation: When using PHP, configure the application so that it does not use register_globals. During implementation, develop the application so that it does not rely on this feature, but be wary of implementing a register_globals emulation that is subject to weaknesses such as CWE-95, CWE-621, and similar issues.

CWE-59: Improper Link Resolution Before File Access ('Link Following')

The product attempts to access a file based on the filename, but it does not properly prevent that filename from identifying a link or shortcut that resolves to an unintended resource.