PenPot MCP REPL server binds to 0.0.0.0 with unauthenticated /execute endpoint — RCE

Summary The MCP module's ReplServer binds to all interfaces 0.0.0.0:4403 and exposes a /execute endpoint that runs arbitrary code with zero authentication. Anyone on the network can POST JavaScript and it runs on the server. The main PenpotMcpServer was partially fixed for a similar binding issue...

FPDI: Memory Exhaustion and Endless Loop in FPDI leads to Denial of Service

Impact This is a significant Denial of Service DoS vulnerability. Any application that uses FPDI to process user-supplied PDF files is at risk. An attacker can upload a small, malicious PDF file that will cause the server-side script to crash due to memory exhaustion or a script time-out. Repeate...

Wire: skipGroup() missing negative-length check allows 10-byte payload to crash any Wire-decoding service

CVE-2026-45799 Maintainer summary Wire's protobuf group-skipping logic did not reject negative lengths before skipping a length-delimited field inside a group. A crafted protobuf payload could cause Wire to throw an unchecked runtime exception during decoding instead of the documented IOException...

Coder: Unauthenticated SSRF via Azure Instance Identity Endpoint

Summary Unauthenticated semi-blind Server-Side Request Forgery SSRF via the Azure instance identity endpoint POST /api/v2/workspaceagents/azure-instance-identity. An external attacker can force the Coder server to issue HTTP GET requests to arbitrary internal or external hosts by submitting a...

HAX CMS: Denial of Service using Malicious Import Request

Summary The HAX CMS NodeJS application crashes when an authenticated attacker sends a specially crafted site creation request to the createSite endpoint. A single request is sufficient to take the entire application offline, requiring a manual server restart to restore service. Details The...

OpenMcdf: Uncatchable infinite loop in DirectoryTree.TryGetDirectoryEntry on crafted CFB directory cycle

Summary The BST name-lookup loop in DirectoryTree.TryGetDirectoryEntry OpenMcdf/DirectoryTree.cs:35-46 walks directory entries by repeatedly calling directories.TryGetSiblingchild, siblingType, validateColor. A crafted CFB file with cyclic Left/Right sibling links among directory entries -...

rust-openssl: Potential out-of-bounds write in `CipherCtxRef::cipher_update_inplace` for AES-KW-PAD ciphers

CipherCtxRef::cipherupdateinplace incorrectly sized output buffers when used with AES key-wrap-with-padding ciphers EVPaes128,192,256wrappad. For a non-multiple-of-8 input, OpenSSL writes up to 7 bytes past the end of the caller's buffer or Vec, producing attacker-controllable heap corruption whe...

Turbo: Login callback CSRF/session fixation

Impact Turborepo's self-hosted login and SSO browser flows did not validate a CSRF state value on the localhost callback. While the CLI was waiting for authentication, a malicious web page could send a request to the local callback server with an attacker-controlled token. If accepted before the...

Turbo: Unexpected local code execution during Yarn Berry detection

Impact Turborepo can be vulnerable to arbitrary code execution when run in untrusted repositories that contain malicious Yarn configuration. In affected versions, package manager detection executed yarn --version from the project directory, which could cause Yarn to load and execute a...

MCP Gateway: Authority-injection and JWT/session bypass via the unauthenticated router hair-pin "router-key" / "mcp-init-host" path

Summary The MCP router extproc exposes an initialize-method code path that, when a request carries an mcp-init-host header, bypasses the gateway JWT session validator and rewrites the upstream :authority header to whatever the caller chooses, gated only by a single shared header value router-key...

Diesel: Command injection in Diesel's implementation of `COPY FROM`/`COPY TO`

Diesel allows users to configure various options for PostgreSQL's COPY FROM and COPY TO statements. These configurations are partially provided as strings or characters. Diesel did not check if any these user-provided options contain a quote character ', which can lead to the injection of...

Diesel: Possible unaligned data access for implementations of `SqliteAggregate`

Diesel allows to register custom aggregate SQL functions for SQLite via the SqliteAggregate interface. To store an instance of the custom aggregate processor Diesel relied on the sqlite3aggregatecontext function provided by sqlite. This function doesn't provide any guarantees about alignment of t...

GitHub CLI: GitHub Actions log output in `gh run view` allows terminal escape sequence injection

Summary A security vulnerability has been identified in GitHub CLI that could allow terminal escape sequence injection when users view GitHub Actions workflow logs using gh run view --log or gh run view --log-failed. Details The vulnerability stems from the way GitHub CLI handles raw Actions log...

Caddy: Remote Admin Authorization Bypass on PKI Endpoints via Prefix-Based Path Matching

AI Disclosure I used an LLM to help review the source code, reason about attack surface, and help draft and refine this report. I manually validated the finding by reproducing it locally, confirming the vulnerable code path, and verifying the HTTP behavior with curl -v. Summary Caddy's remote adm...

Caddy CVE-2026-30852 Fix Bypass

TL;DR CVE-2026-30852 fixed double expansion in varsregexp when the variable key is a placeholder e.g. http.vars.x. The fix does NOT protect literal key names e.g. tenantid. An attacker injects env.AWSSECRETACCESSKEY or file./etc/passwd via a request header → Caddy expands it on the second pass →...

Kong Ingress Controller for Kubernetes (KIC): Cross-namespace TLS Secret Exfiltration in Gateways with GatewayClass missing `konghq.com/gatewayclass-unmanaged: 'true'` annotation

Summary A vulnerability in the Kong Ingress Controller KIC allows for the unauthorized exfiltration of TLS certificates and private keys across Kubernetes namespace boundaries. In "managed" mode where the GatewayClass lacks an unmanaged annotation, the Gateway TLS translator skips critical status...

Kong Ingress Controller for Kubernetes (KIC): Secret-backed plugin configurations leak through non-sensitive diagnostics endpoint

Summary A vulnerability in the Kong Ingress Controller KIC allows for the unauthorized exposure of sensitive plugin credentials through the diagnostics interface. Even when configured to redact sensitive information using --dump-sensitive-config=false, KIC fails to sanitize the Plugins field in...

Bandit: Unauthenticated DoS via chunked request trailers in Bandit HTTP/1 decoder

Summary A worker-pinning denial of service in Bandit's HTTP/1 chunked transfer decoder. Any unauthenticated client that sends a Transfer-Encoding: chunked request whose body ends with a trailer field RFC 9112 §7.1.2 explicitly permits this causes the connection's worker process to spin forever in...

Bandit: Unauthenticated one-shot DoS via `Transfer-Encoding: chunked`

Summary Bandit's HTTP/1 chunked-body reader silently drops the request size cap that the application configures e.g. Plug.Parsers' default 8 MB length: and buffers the entire body in memory before the application sees it. An unauthenticated attacker can crash any Bandit-fronted Phoenix/Plug app...

9router: Unauthenticated Remote Code Execution via unprotected MCP custom plugin routes

Summary 9router exposes two unauthenticated API endpoints that, when chained together, allow any network-adjacent attacker to execute arbitrary OS commands as the user running the 9router process — with zero prerequisites and no credentials required. The vulnerability exists because the Next.js...

Kopia: RCE via SSH ProxyCommand Injection

Summary Kopia's HTTP server, when started with --without-password , accepts unauthenticated requests to /api/v1/repo/exists. The handler forwards an attacker-supplied storage configuration to blob.NewStorage. For SFTP backends with externalSSH: true, that path constructs a process command line by...

Jaspersoft Reports: Java Deserialization Vulnerability Lleads to Remote Code Execution (RCE)

Java Deserialisation Vulnerability in Jaspersoft Reports Library leads to Remote Code Execution RCE, potentially allowing code execution on the affected system...

APScheduler's JSONSerializer and CBORSerializer are vulnerable to Remote Code Execution (RCE) via Insecure Deserialization

The JSONSerializer and CBORSerializer in APScheduler all versions including 3.10.x and 4.0.0a5 are vulnerable to Remote Code Execution RCE via Insecure Deserialization. The unmarshalobject function allows for arbitrary class instantiation and state injection by dynamically importing modules and...

BillaBear is Vulnerable to SQL Injection in the EventRepository

BillaBear all versions prior to Jan 2026 contains a SQL Injection vulnerability in the EventRepository. User-controlled input from metric filter names and aggregation properties is directly interpolated into SQL queries using sprintf without proper sanitization or identifier quoting. Although...

Apify Model Context Protocol (MCP) server: Domain Allowlist Bypass in fetch-apify-docs via String Prefix Matching

Summary The fetch-apify-docs tool validates URLs against a domain allowlist using String.startsWith instead of proper URL hostname comparison. This allows bypass via attacker-controlled subdomains e.g., https://docs.apify.com.evil.com/, enabling the tool to fetch and return arbitrary web content ...

Budibase: Unrestricted Upload of File with Dangerous Type

Summary The file upload endpoint POST /api/attachments/process does not enforce active-content restrictions for authenticated users. The checks for dangerous file extensions html, svg, js, php, etc. are conditionally wrapped inside if isPublicUser or if isPublicUser || !env.SELFHOSTED, meaning an...

Budibase: Missing Cache Invalidation on Public API Role Unassignment Allows Revoked Users to Retain Privileges for Up to 1 Hour

Summary The public API role unassignment endpoint POST /api/public/v1/roles/unassign updates user documents in CouchDB but does not invalidate the corresponding Redis user cache entries. Because the authentication middleware resolves user identity and permissions from this cache TTL: 3600 seconds...

AVideo: Unauthenticated Arbitrary Image Read via Path Traversal in `view/img/image404Raw.php`

Summary The endpoint requires no authentication. An unauthenticated remote attacker can read arbitrary image files anywhere on disk that the PHP user can open — including private user-profile photos that the application's normal serving wrappers gate behind ACLs, admin-uploaded thumbnails,...

protobufjs: Denial of Service via unbounded recursive JSON descriptor expansion

Summary protobufjs could recurse without a depth limit while expanding nested JSON descriptors through Root.fromJSON and Namespace.addJSON. A crafted JSON descriptor with deeply nested namespace definitions could cause the JavaScript call stack to be exhausted during descriptor loading. Impact An...

libcrux-ml-dsa: Signature Verification on AVX2 Platforms Mishandles Edge Case

The AVX2 implementation of ML-DSA verification incorrectly implemented the usehint function, mishandling an edge case that should lead to signature rejection. Impact An attacker could make the ML-DSA verifier accept a crafted invalid signature under a maliciously generated verification key, if th...

libcrux: Potential Panic on Overlong Ciphertext Buffer

An application that passes in a ciphertext buffer of length greater than ptxt.len + TAGLEN to libcruxchacha20poly1305::encrypt or libcruxchacha20poly1305::xchacha20poly1305::encrypt would experience a panic. Impact An application where the length of the ciphertext buffer is under attacker control...

Envoy AI Proxy - MCP Message Smuggling Vulnerability

Envoy AI Gateway was found to be affected by a protocol parser differential vulnerability due to improper implementation of the JSON-RPC 2.0 specification. Such differential causes a MCP message alteration, potentially causing a bypass of security controls in a multi-layered architecture. Accordi...

Github Actions issued GITHUB_TOKEN disclosure in GitHub Actions logs

Summary Composer leaks the full contents of tokens configured as GitHub OAuth tokens if they do not match Composer's expected format for such tokens to stderr. GitHub has introduced a new format for GitHub Actions GITHUBTOKEN values. These tokens are validated in the same way by Composer on GitHu...

n8n: Credential exfiltration via Allowed HTTP Request Domains Bypass

Impact The POST /rest/dynamic-node-parameters/options endpoint allowed any authenticated user to cause the n8n server to issue HTTP requests including credentials bypassing the intended restrictions on which hosts could be contacted for that credential Allowed HTTP Request Domains. The user neede...

n8n: Legacy ExecuteWorkflow Node Bypassed File Path Restrictions

Impact The ExecuteWorkflow node's localFile source option read workflow files from disk without applying checks enforced by other file-reading nodes. An authenticated user with permission to create or modify workflows could supply an arbitrary file path via the REST API, bypassing the...

Strawberry GraphQL: Default GraphiQL may expose HTTP headers in URLs

Summary Strawberry's bundled GraphiQL template wrote values from the GraphiQL headers editor into the browser URL query string. If a user entered a sensitive header, such as Authorization: Bearer , the value could become visible in browser history, copied links, and server/proxy/CDN access logs...

Argo CD: Stored XSS in application link annotations enables developer-to-admin privilege escalation

Summary A user with application write access developer role can set link.argocd.argoproj.io/ annotations on any ArgoCD Application. These annotation values are rendered in the Summary tab's URLs section as elements without URL validation. Using the pipe-separator trick Display Text |...

Argo CD: Kubernetes Secret Extraction via ArgoCD ServerSideDiff via sensitive annotations

Summary The original fix for GHSA-3v3m-wc6v-x4x3 is incomplete. argocd app diff --server-side-diff can still expose Kubernetes Secret values embedded in the kubectl.kubernetes.io/last-applied-configuration annotation. The prior fix masks top-level Secret data in ServerSideDiff responses, but it...

Mailpit: Unauthenticated remote memory-exhaustion DoS via unlimited SMTP DATA and /api/v1/send body sizes

Summary The Mailpit SMTP server has a Server.MaxSize int field that controls the maximum allowed DATA payload size, but the field is never assigned anywhere outside test code, leaving it at Go's zero value 0 ⇒ "no limit". The same applies to the HTTP /api/v1/send endpoint, whose request body is...

Mailpit: Concurrent map read & write in proxy CSS rewriter - remote unauth crash (fatal error: concurrent map read and map write)

Summary The screenshot/print proxy /proxy?data=… maintains a package-level assets mapstringMessageAssets cache, but reads the map without holding assetsMutex while a long-running cleanup goroutine and re-entrant CSS-rewriting code path concurrently write to it under the lock. When the...

Mailpit: Path traversal & arbitrary file write in mailpit dump --http via attacker-controlled message IDs

Summary The mailpit dump --http sub-command downloads every message from a remote Mailpit instance and writes each one as .eml inside the user-supplied output directory. The message ID field is taken verbatim from the JSON response of the remote server and concatenated into the output path with...

Mailpit has an incomplete fix for GHSA-6jxm: HTML check still permits SSRF to private/loopback/IMDS via missing IP-filter dialer

Summary The fix for GHSA-6jxm-fv7w-rw5j CVE-2026-23845, "Server-Side Request Forgery SSRF via HTML Check API", shipped in mailpit v1.28.3, hardened internal/htmlcheck/css.go::downloadCSSToBytes with a 5MB size cap, a text/css content-type check, login-info stripping in isValidURL, and an opt-in...

Caddy: Remote Admin Authorization Bypass in `/config` API via Array Index Normalization

This report is not about a normal textual prefix-expansion case. The issue here is that the authorization layer and the /config traversal layer do not agree on what object the path refers to. In this case, a path authorized for one config object is accepted, but then resolves to a different confi...

Nuxt: Dev server exposes built source over LAN to malicious sites (incomplete fix for GHSA-4gf7-ff8x-hq99)

Summary This is an incomplete fix for GHSA-4gf7-ff8x-hq99. Source code may be stolen during dev when using the webpack / rspack builder if the dev server is bound to a non-loopback address e.g. nuxt dev --host and the developer opens a malicious site on the same network. Details The fix for...

Nuxt: Reflected XSS in `navigateTo()` external redirect

Summary navigateTo with external: true generates a server-side HTML redirect body containing a tag. The destination URL is only sanitized by replacing " with %22, leaving , &, and ' unencoded. An attacker who can influence the URL passed to navigateTourl, external: true can break out of the...

ORAS Java: Path traversal in pullArtifact via attacker-controlled org.opencontainers.image.title annotation

Summary The pullArtifact methods in Registry and OCILayout use the org.opencontainers.image.title annotation from a pulled manifest as a filename, resolving it against the caller supplied output directory without normalization or a containment check. A manifest publisher can set this annotation t...

auth-fetch-mcp: SSRF and disk exfiltration via unvalidated auth_fetch and download_media URLs

SSRF + disk-exfil in downloadmedia and authfetch tools — ymw0407/auth-fetch-mcp Severity The downloadmedia and authfetch MCP tools accept arbitrary URLs and reach them as the MCP server process, with downloadmedia additionally persisting the fetched response body to a user-controlled output...

Malicious code in guardrails-ai 0.10.1 (supply chain compromise)

Impact On May 11, 2026 at approximately 6:00 PM Pacific, an attacker published a malicious version of guardrails-ai 0.10.1 to PyPI. Affected: any user who installed guardrails-ai==0.10.1 from PyPI on May 11, 2026. Security researchers identified the malicious package within approximately 2 hours ...

fabric-chaincode-java: TLS Private Key Password Disclosed in INFO Startup Logs in Chaincode-as-a-Service Mode

When chaincode is deployed in chaincode-as-a-service mode with TLS enabled, the chaincode server INFO level logging includes the TLS private key password in plaintext. An attacker with access to the chaincode server logs could recover the TLS private key password. If the attacker can also obtain...

MCP Registry: OCI validator skips ownership check on upstream rate limits

OCI ownership validation fails open on upstream rate limits, allowing attacker to claim arbitrary public OCI images under their own namespace Severity: Low re-scored post-triage; see Maintainer triage note below Affected: modelcontextprotocol/registry main branch at commit fe0cb3b current HEAD as...