May 7, 2026·4 min read·25 visits
A logic flaw in the Bandit HTTP server allows attackers to forge the connection scheme. Sending an absolute-form HTTP request over a plaintext connection causes the server to treat the request as HTTPS, bypassing SSL redirects and leaking secure cookies.
The Bandit HTTP server prior to version 1.11.0 contains a transport-state spoofing vulnerability. The application incorrectly prioritizes the client-supplied URI scheme over the verified transport-layer encryption status. This allows unauthenticated attackers to spoof the connection state as secure (HTTPS) over a plaintext connection, bypassing security middleware and exposing secure cookies.
Bandit is an HTTP server written in Elixir, responsible for accepting client connections and generating a Plug.Conn struct for downstream application logic. The conn.scheme attribute communicates the security context of the connection to the application.
CVE-2026-39807 is a transport-state spoofing vulnerability categorized as CWE-807 (Reliance on Untrusted Inputs in a Security Decision). The server incorrectly prioritizes the URI scheme supplied by the client over the verified state of the underlying transport layer.
This architecture failure permits unencrypted HTTP connections to present themselves as secure HTTPS connections within the application context. Downstream middleware relies on this spoofed context to make security decisions, resulting in multiple logic bypasses.
The vulnerability originates from a regression introduced in June 2023 via commit ff2f8293. The function Bandit.Pipeline.determine_scheme/2 in lib/bandit/pipeline.ex processes incoming request metadata to populate the Plug.Conn struct.
The function implementation evaluated a tuple containing both the transport adapter's secure? boolean and the client-supplied scheme. The final case match in this function blindly extracted the client-supplied scheme.
By matching on the client-supplied data irrespective of the secure? flag, the pipeline elevated an untrusted header to an authoritative state property. The transport layer's native encryption state was explicitly discarded.
The vulnerable code in lib/bandit/pipeline.ex demonstrates the flaw in prioritization. The third match arm returns the client-provided scheme regardless of the physical secure? transport status.
# lib/bandit/pipeline.ex (Vulnerable)
defp determine_scheme({secure?, _, _, _}, {scheme, _, _, _}) do
case {scheme, secure?} do
{nil, true} -> {:ok, "https"}
{nil, false} -> {:ok, "http"}
{scheme, _} -> {:ok, scheme} # Vulnerability: Accepts client scheme
end
endCommit 45feea20dea8af7ffd7245271107b695c040e667 resolves this behavior by entirely removing the client scheme parameter from the evaluation function.
# lib/bandit/pipeline.ex (Fixed)
@spec determine_scheme(boolean()) :: String.t()
defp determine_scheme(true), do: "https"
defp determine_scheme(false), do: "http"The patched version guarantees that the application scheme strictly mirrors the transport layer's physical configuration, eliminating the influence of client headers.
Exploitation requires network access to a plaintext port served by Bandit. The attacker initiates a standard TCP connection without TLS negotiation.
The attacker issues an HTTP/1.1 absolute-form request target or an HTTP/2 request with a manipulated :scheme pseudo-header. The absolute-form syntax explicitly specifies the https scheme in the request line.
GET https://target-server.internal/restricted-endpoint HTTP/1.1
Host: target-server.internalBandit processes this payload and populates the Plug.Conn struct with scheme: :https. The connection proceeds through the application pipeline possessing a falsified security context.
The spoofed scheme systematically bypasses Elixir security middleware. The Plug.SSL module validates conn.scheme to enforce HTTP-to-HTTPS redirection. Under exploitation, Plug.SSL observes the spoofed :https value and skips the redirection block.
Authentication mechanisms that issue session cookies with the secure: true flag will transmit these tokens over the unencrypted connection. The application operates under the false assumption that it is writing to an encrypted socket.
Network interceptors or adversaries positioned on the local network segment can capture these secure session tokens. CSRF prevention mechanisms and SameSite cookie validations that perform origin scheme matching are similarly bypassed.
The primary remediation is upgrading the bandit dependency to version 1.11.0. This version contains the logic patch that strictly ties the scheme to the underlying transport security state.
Administrators who cannot immediately patch must deploy compensating controls. Reverse proxies placed in front of Bandit can rewrite the HTTP request lines to strip absolute-form targets.
Disabling all plaintext listeners on the Bandit server eliminates the attack vector entirely. When Bandit only listens on TLS-enabled sockets, the secure? flag is permanently true, nullifying the scheme spoofing objective.
CVSS:4.0/AV:N/AC:L/AT:P/PR:N/UI:N/VC:L/VI:L/VA:N/SC:N/SI:N/SA:N| Product | Affected Versions | Fixed Version |
|---|---|---|
bandit mtrudel | >= 1.0.0, < 1.11.0 | 1.11.0 |
| Attribute | Detail |
|---|---|
| CWE ID | CWE-807 |
| Attack Vector | Network |
| CVSS Score | 6.3 |
| EPSS Score | 0.00018 |
| Impact | Security feature bypass and confidentiality loss |
| Exploit Status | Proof of Concept |
| KEV Status | Not Listed |
The system makes a security decision based on input from an untrusted source, which can be modified by an attacker.
The @jhb.software/payload-cloudinary-plugin exposes an endpoint that performs unvalidated cryptographic signing of Cloudinary API parameters, allowing authenticated users with minimal privileges to forge valid signatures for arbitrary actions. This flaw allows attackers to overwrite remote storage assets, execute unauthorized file uploads, alter asset visibility parameters, trigger SSRF webhooks, and perform directory traversal within Cloudinary repositories.
A Server-Side Request Forgery (SSRF) and Bearer Token Exfiltration vulnerability exists in the @merill/lokka (Lokka) Model Context Protocol (MCP) server prior to version 2.1.2. The server constructed Azure Resource Manager request URLs by concatenating user-controlled path parameters directly into destination request strings. By injecting authority-redefinition characters, an attacker can manipulate URL parsing to execute a host-escape attack, forcing the server to send high-privilege Azure Resource Manager (ARM) Bearer tokens to an external attacker-controlled host. This allows complete administrative access to the associated Azure subscriptions.
A directory traversal and symlink following vulnerability exists in Pydantic Settings when using the NestedSecretsSettingsSource with nested subdirectory lookups enabled. An attacker capable of writing to the secrets directory can bypass size limitations, read arbitrary host files, or cause a denial-of-service condition via cyclic symlinks.
A Server-Side Request Forgery (SSRF) vulnerability exists in SurrealDB's Identity & Access Management (IAM) module prior to version 3.1.5. When configuring JSON Web Key Set (JWKS) URLs for token verification, the remote fetcher follows HTTP redirects by default without validating redirect targets against configured network capabilities. This allows high-privileged users to bypass network access limits and perform blind port scanning of internal network resources.
A local file disclosure vulnerability exists in SurrealDB's full-text search capabilities, allowing authenticated users with database EDITOR or OWNER roles to read arbitrary files from the host system filesystem. This occurs by abusing the mapper() filter inside a DEFINE ANALYZER statement to point to system files.
SurrealDB versions 3.0.0 through 3.1.4 contain an information exposure vulnerability (CWE-203) where the query planner optimizes sorted queries using indexes on fields with field-level SELECT restrictions. Because the query planner performs index-based sorting before enforcing permission-based redaction, unauthorized users can observe the physical order of returned rows to deduce the relative values of protected fields.