CVE Database

OpenClaw (npm package 'openclaw') versions >= 2026.3.28 and < 2026.8.1 allow an untrusted workspace .env file to set the CLOUDSDK_PYTHON_ARGS environment variable. When an operator starts OpenClaw in attacker-controlled workspace content and then runs the Gmail setup flow, that value is inherited when gcloud is launched, and the gcloud launcher passes it as arguments to the trusted Python interpreter. A crafted CLOUDSDK_PYTHON_ARGS value can therefore cause Python to execute attacker-supplied code with the OpenClaw host user's permissions, allowing credentials to be read, files to be modified, or other processes to be started. This issue is fixed in version 2026.8.1; as a workaround, run Gmail setup only from trusted workspaces and clear inherited CLOUDSDK_* variables beforehand.

OpenClaw versions before 2026.8.1 fail to properly restrict access to operator command cron jobs, allowing model-visible agent callers to read and execute ownerless command jobs. Attackers can inspect stored environment variables and force-run disabled or unscheduled command jobs to access secrets and execute operator-authored commands.

OpenClaw is an agent gateway distributed as the npm package 'openclaw'. In versions >= 2026.4.5 and < 2026.8.1, the Gateway validated a single DNS resolution result for a configured remote Chrome DevTools Protocol (CDP) hostname, but the raw WebSocket and Playwright transports performed a later, independent DNS resolution, discarding the DNS pinning enforced at validation time. An attacker who controls an approved CDP hostname or its DNS answers can exploit this check-then-use gap via DNS rebinding to make the Gateway connect to a loopback, private, link-local, cloud metadata, or other SSRF-policy-denied address. The impact depends on the remote CDP configuration, DNS timing, and the services reachable from the Gateway host. The issue is fixed in 2026.8.1; as a mitigation, disable hostname-based remote CDP endpoints or restrict them to trusted, stable infrastructure.

OpenClaw (npm package 'openclaw') versions >= 2026.3.22 and < 2026.8.1 contain an approval-bypass flaw in the exec approval policy: the policy could trust a command-running wrapper without inspecting the command carried in its arguments. After an operator allowlisted or permanently approved a benign wrapper invocation, a later agent turn could substitute an arbitrary inner command and execute it with the OpenClaw process's host privileges without a further approval prompt. Exploitation requires the relevant wrapper to resolve on the host and a prior operator decision allowing that wrapper. Transparent shell carriers and opaque utilities such as process monitors, tracers, namespace tools, and proxy wrappers were affected through related trust-resolution gaps. Fixed in 2026.8.1.

OpenClaw versions before 2026.8.1 contain an authorization bypass vulnerability where Allow Always approvals for exact commands persist as path-only grants on macOS and Linux. Attackers can reuse the same executable with different arguments to execute commands without triggering new approval prompts, potentially accessing files or internal services.

OpenClaw versions before 2026.8.1 contain a command parser vulnerability where escaped newlines confuse exec allowlist parsing, allowing hidden commands to execute. Attackers can craft input with escaped newlines to bypass allowlist validation and execute additional commands without expected authorization prompts.

OpenClaw versions before 2026.8.1 contain a resource exhaustion vulnerability in the Gateway listener that allows unauthenticated clients to retain response sockets by sending WebSocket upgrade requests without matching connection semantics. Attackers can repeatedly send malformed upgrade requests to exhaust listener resources and cause denial of service without consuming the WebSocket pre-auth connection budget.

OpenClaw versions before 2026.8.1 contain an authorization bypass vulnerability in skill tool dispatch that fails to carry the sender's owner status. Non-owner senders authorized to invoke skill commands can access owner-only tools and server credentials reserved for owners.

OpenClaw is an npm-distributed gateway application. In versions >= 2026.7.1 and < 2026.8.1, Synology Chat attachment delivery could lose DNS pinning: the Gateway validated a single DNS result for a supplied file URL but then passed the original hostname to the Synology NAS, where it could resolve to a different destination. When attachment delivery accepted a remotely influenced hostname, an attacker could use DNS rebinding to make the NAS fetch a private or otherwise policy-denied resource and return its contents to the addressed conversation (server-side request forgery). Practical impact depends on NAS routing, resolver behavior, and the response available at the private destination. The issue is fixed in 2026.8.1; as a workaround, disable remote URL attachment forwarding in Synology Chat.

OpenClaw (npm package 'openclaw') before 2026.8.1 does not correctly enforce per-chat tool policies for Codex app-server runtime tools. A conversation-level tools.allow rule filtered OpenClaw tools but did not restrict the shell, process, file, and patch tools owned by the Codex runtime. When a lower-trust conversation was assigned to a Codex runtime and restricted with a per-chat tool allowlist, a participant able to trigger that agent could still reach native command and file tools, bypassing the configured allowlist. The practical impact depends on the runtime's host permissions and sandbox configuration. The issue is fixed in 2026.8.1.

OpenClaw for iOS versions >= 2026.7.1 and < 2026.8.11 do not enforce saved Gateway TLS pins in the Control UI. While native connections enforced the saved Gateway fingerprint, the authenticated Terminal and session Dashboard WebViews omitted it. If a user had accepted a Gateway fingerprint, an attacker able to redirect the same host and port and present a different certificate that is accepted by iOS system trust can serve a replacement Control UI page; opening the Terminal or a session Dashboard then allows that page to read the injected Gateway token or password. The stolen credential can grant operator access, including reading sensitive Gateway state and invoking host-capable tools. This issue is fixed in 2026.8.11.

openclaw's @openclaw/voice-call package before 2026.8.1 launches the configured agent for classic inbound voice calls without propagating the caller's identity or non-owner status. As a result, owner-only tool filtering can fail open and expose the agent's normal tool authority to a remote caller. A caller who is admitted by the configured inbound-call policy (open, pairing, or allowlist) on a deployment with inbound calling enabled can therefore drive tools intended for the trusted owner, potentially reading data, modifying files, executing commands, or controlling connected services depending on the agent's configuration. The issue is fixed in 2026.8.1.

OpenClaw (npm package openclaw) before 2026.8.1 could include deterministic hashes computed over the original, unredacted configuration in redacted configuration responses. When the Gateway password had low entropy and the remaining configuration values were reconstructable, these hashes acted as offline password verifiers: a caller able to obtain the redacted configuration (for example via config.get) could test password candidates offline without going through the rate-limited Gateway authentication path. Recovering the password could grant the documented shared-secret operator authority. Secret references were not affected in the same way. The issue is fixed in 2026.8.1.

OpenClaw's Matrix integration (npm package @openclaw/matrix) versions >= 2026.2.2 and < 2026.8.1 lowercase complete Matrix user IDs — including historical localparts and the case-sensitive server-name portion — when deriving the OpenClaw authorization identity. As a result, distinct authenticated Matrix accounts can normalize to the same authorization identity. A Matrix participant controlling a colliding account identifier (a protocol-valid identifier that differs from the configured one only by characters OpenClaw case/Unicode folds; display-name matching is not required) can inherit allowlist, owner-command, exec-approval, or plugin-approval authority configured for another account. The issue is fixed in 2026.8.1.

OpenClaw (npm package 'openclaw') versions >= 2026.4.5 and < 2026.8.1 can lose the originating requester's restrictions and untrusted provenance when session-derived text is persisted to session memory. In deployments where session-memory capture and dreaming are enabled, a restricted external sender whose messages are admitted with limited tools can persist instructions that are later supplied to an unattended background (dreaming) agent holding broader file and command capabilities, allowing actions beyond the authority of the original turn and affecting files, commands, or services available to that agent. Exploitation requires the content to be captured, selected for later processing, and followed by the model. The issue is fixed in 2026.8.1.

@openclaw/whatsapp (npm) before 2026.8.1 exposes the WhatsApp login tool through the generic channel-tool path without preserving the originating sender's owner status, so the owner-only tool boundary is not enforced. An admitted non-owner sender able to steer the tool can request a forced login and receive a new QR code for a configured account, disconnecting the Gateway's WhatsApp account and causing loss of availability; full account relinking additionally requires the attacker to scan the returned QR code with another phone. The issue affects the owner-only tool boundary rather than WhatsApp transport authentication. Fixed in 2026.8.1.

OpenClaw versions before 2026.8.1 fail to bind working directory context to reusable exec approvals, allowing approved commands to execute in different directories. Attackers with an allow-always approval can reuse it to run the same command against unreviewed files or repositories with materially different effects.

Laranode versions before 1.2.1 contain a path traversal vulnerability in the POST /filemanager/upload-file endpoint that allows authenticated users to write arbitrary files outside their home directory. Attackers can supply directory traversal sequences in the path parameter to write PHP files into other tenants' web roots and execute code as those tenants.

Ghidra versions through 12.1.4 contain a stack-based out-of-bounds write vulnerability in the decompiler's leftshift128 function when processing negative shift amounts from p-code. Attackers can craft malicious binaries with specific instruction sequences that trigger the overflow when decompiled, corrupting memory and potentially achieving code execution.

Horilla is an HR and CRM software. Prior to 2.0.0, HorillaListView.export_data in horilla_views/generic/cbv/views.py accepts an authenticated user's columns POST parameter, takes field_tuple[1], interpolates it into dynamic_fn_str as Python source, and passes the generated function definition to exec(). A crafted string that remains valid under ast.literal_eval can inject Python syntax into a default argument evaluated during function definition, allowing arbitrary operating-system commands to execute with the application process privileges, including root privileges in the shipped Docker image. This issue is fixed in version 2.0.0.

Actual is a local-first personal finance tool. Prior to 26.7.0, Actual Sync Server's CORS proxy is intended to let authenticated users fetch resources only from repositories listed in the official plugin allowlist. When `ACTUAL_GITHUB_TOKEN` is configured, the proxy automatically attaches the server's GitHub token to GitHub requests. The GitHub API allowlist check uses a raw `startsWith()` prefix test for `/repos/{owner}/{repo}` without requiring a path boundary after the repository name. If an allowlisted public plugin repository is `https://github.com/acme/plugin`, the proxy also accepts GitHub API URLs. Those URLs are outside the allowlisted repository but still pass because their API path starts with `/repos/acme/plugin`. The proxy then forwards the request with the server's `ACTUAL_GITHUB_TOKEN`, allowing any authenticated Actual user to read private GitHub resources reachable by that token. Version 26.7.0 fixes the issue.

InvoicePlane is a self-hosted open source application for managing invoices, clients, and payments. Prior to 1.7.2, InvoicePlane fails to revoke administrative privileges after a role downgrade because Admin_Controller trusts the user_type snapshot stored in an existing session instead of revalidating ip_users.user_type. When one administrator downgrades another account, the target's active session continues to authorize administrative requests. The downgraded user can use Users::form() to set user_type back to 1, restoring the database role and making the privilege escalation persistent. This vulnerability is fixed in 1.7.2.

Horilla is an HR and CRM software. Prior to 1.6.0, the search parameter at /employee/employee-filter-view is reflected by jQuery .html() in employee/templates/employee_nav.html without HTML neutralization. An external attacker can craft and deliver a link that causes JavaScript to execute when an authenticated employee or administrator reaches the employee filter, allowing access to browser-visible session data and actions with the victim's application privileges. This issue is fixed in version 1.6.0.

gitoxide gix-fs before 0.23.0 contains a path validation bypass vulnerability in the worktree checkout mechanism that allows attackers to escape the worktree directory via symlink manipulation. During forced checkout with overwrite_existing enabled, attackers can craft malicious repository trees where symlink entries replace validated directories, causing subsequent files to be written outside the worktree through the symlink for code execution or file manipulation.

cleanup_xml_node() in the SOAP XML parser recurses once per XML nesting level with no depth limit. An unauthenticated attacker can post a SOAP request containing tens of thousands of nested elements to any SoapServer endpoint, exhaust the stack and crash the process. The same unbounded recursion exists in the SOAP value decoder and in the WSDL node search helper.

SCBE-AETHERMOORE is a geometric AI governance and evaluation framework. Starting in version 4.0.2 and prior to version 4.2.1, the AetherBrowser API server (`scripts/aetherbrowser/api_server.py`) exposes the `POST /api/ops/check-email` endpoint without any authentication. Any remote attacker can call this endpoint and trigger execution of the `email_reader.py` subprocess, which connects to configured ProtonMail or Gmail accounts via IMAP and returns email metadata (sender, subject, body snippet) in the JSON response. The server binds to `0.0.0.0:8100` by default with CORS set to `allow_origins=["*"]`, making it reachable from any network or browser origin. Version 4.2.1 patches the issue.

Esri LERC is an open-source image or raster format which supports rapid encoding and decoding for any pixel type. A Heap based Out-of-Bounds Write via Integer Overflow in LERC versions 4.1.0 and earlier may allow a remote, unauthenticated attacker who can pass specifically crafted attacker controlled imagery to an application that uses LERC to crash the application, leading to a denial of service.

MediaFlow Proxy through 2.4.9 contains a server-side request forgery vulnerability in the /proxy routes due to missing and incomplete destination validation in the d query parameter. Remote attackers can supply arbitrary internal URLs including loopback and cloud metadata endpoints to read full responses from the proxy server.

Zoraxy versions 3.2.3 through 3.3.4 fail to properly parse IPv6 addresses in the RemoteAddr field when setting forwarded headers. Unauthenticated attackers connecting over IPv6 can supply arbitrary X-Forwarded-For values to spoof their source IP and bypass authorization provider IP-based access controls.

GestSup versions before 3.2.62 contain a remote code execution vulnerability in the basic IMAP connector's attachment handling that fails to skip blocked file extensions. Unauthenticated attackers can send emails with PHP attachments to monitored mailboxes, which are written to the web-accessible upload/ticket directory and executed when accessed.