OpenBao is an open source identity-based secrets management system. Prior to version 2.5.3, when OpenBao revoked privileges on a role in the PostgreSQL database secrets engine, OpenBao failed to use proper database quoting on schema names provided by PostgreSQL. This could lead to role revocation failures, or more rarely, SQL injection as the management user. This vulnerability was original from HashiCorp Vault. The vulnerability is addressed in v2.5.3. As a workaround, audit table schemas and ensure database users cannot create new schemas and grant privileges on them.

Claude Code is an agentic coding tool. Prior to version 2.1.64, Claude Code's sandbox did not prevent sandboxed processes from creating symlinks pointing to locations outside the workspace. When Claude Code subsequently wrote to a path within such a symlink, its unsandboxed process followed the symlink and wrote to the target location outside the workspace without prompting the user for confirmation. This allowed a sandbox escape where neither the sandboxed command nor the unsandboxed app could independently write outside the workspace, but their combination could write to arbitrary locations, potentially leading to code execution outside the sandbox. Reliably exploiting this required the ability to add untrusted content into a Claude Code context window to trigger sandboxed code execution via prompt injection. Users on standard Claude Code auto-update have received this fix automatically. Users performing manual updates are advised to update to version 2.1.64 or later.

OpenBao is an open source identity-based secrets management system. Prior to version 2.5.3, `ExtractPluginFromImage()` in OpenBao's OCI plugin downloader extracts a plugin binary from a container image by streaming decompressed tar data via `io.Copy` with no upper bound on the number of bytes written. An attacker who controls or compromises the OCI registry referenced in the victim's configuration can serve a crafted image containing a decompression bomb that decompresses to an arbitrarily large file. The SHA256 integrity check occurs after the full file is written to disk, meaning the hash mismatch is detected only after the damage (disk exhaustion) has already occurred. This allow the attacker to replace **legit plugin image** with no need to change its signature. Version 2.5.3 contains a patch.

OpenBao is an open source identity-based secrets management system. Prior to version 2.5.3, OpenBao's Certificate authentication method, when a token renewal is requested and `disable_binding=true` is set, attempts to verify the current request's presented mTLS certificate matches the original. Token renewals for other authentication methods do not require any supplied login information. Due to incorrect matching, the certificate authentication method would allow renewal of tokens for which the attacker had a sibling certificate+key signed by the same CA, but which did not necessarily match the original role or the originally supplied certificate. This implies an attacker could still authenticate to OpenBao in a similar scope, however, token renewal implies that an attacker may be able to extend the lifetime of dynamic leases held by the original token. This attack requires knowledge of either the original token or its accessor. This vulnerability is original from HashiCorp Vault. This is addressed in v2.5.3. As a workaround, ensure privileged roles are tightly scoped to single certificates.

Neko is a a self-hosted virtual browser that runs in Docker and uses WebRTC In versions 3.0.0 through 3.0.10 and 3.1.0 through 3.1.1, any authenticated user can immediately obtain full administrative control of the entire Neko instance (member management, room settings, broadcast control, session termination, etc.). This results in a complete compromise of the instance. The vulnerability has been patched in v3.0.11 and v3.1.2. If upgrading is not immediately possible, the following mitigations can reduce risk: Restrict access to trusted users only (avoid granting accounts to untrusted parties); ensure all user passwords are strong and only shared with trusted individuals; run the instance only when needed; avoid leaving it continuously exposed; place the instance behind authentication layers such as a reverse proxy with additional access controls; disable or restrict access to the /api/profile endpoint if feasible; and/or monitor for suspicious privilege changes or unexpected administrative actions. Note that these are temporary mitigations and do not fully eliminate the vulnerability. Upgrading is strongly recommended.

The nbconvert tool, jupyter nbconvert, converts Jupyter notebooks to various other formats via Jinja templates. In versions 6.5 through 7.17.0, when `HTMLExporter.embed_images=True`, nbconvert's markdown renderer allows arbitrary file read via path traversal in image references. A malicious notebook can exfiltrate sensitive files from the conversion host by embedding them as base64 data URIs in the output HTML. nbconvert 7.17.1 contains a fix. As a workaround, do not enable `HTMLExporter.embed_images`; it is not enabled by default.

The nbconvert tool, jupyter nbconvert, converts Jupyter notebooks to various other formats via Jinja templates. Versions 6.5 through 7.17.0 allow arbitrary file writes to locations outside the intended output directory when processing notebooks containing crafted cell attachment filenames. The `ExtractAttachmentsPreprocessor` passes attachment filenames directly to the filesystem without sanitization, enabling path traversal attacks. This vulnerability provides complete control over both the destination path and file extension. Version 7.17.1 contains a patch.

Signal K Server is a server application that runs on a central hub in a boat. Versions prior to 2.25.0 are vulnerable to an unauthenticated Regular Expression Denial of Service (ReDoS) attack within the WebSocket subscription handling logic. By injecting unescaped regex metacharacters into the `context` parameter of a stream subscription, an attacker can force the server's Node.js event loop into a catastrophic backtracking loop when evaluating long string identifiers (like the server's self UUID). This results in a total Denial of Service (DoS) where the server CPU spikes to 100% and becomes completely unresponsive to further API or socket requests. Version 2.25.0 contains a fix.

OpenClaw before 2026.3.31 contains a resource consumption vulnerability in Telegram audio preflight transcription that allows unauthorized group senders to trigger transcription processing. Attackers can exploit insufficient allowlist enforcement to cause resource or billing consumption by initiating audio preflight operations before authorization checks are applied.

OpenClaw before 2026.3.31 contains an environment variable override vulnerability in host exec policy that fails to properly enforce proxy, TLS, Docker, and Git TLS controls. Attackers can bypass security controls by overriding environment variables to circumvent proxy settings, TLS verification, Docker restrictions, and Git TLS enforcement.

OpenClaw before 2026.3.31 contains a sandbox bypass vulnerability allowing attackers to escalate privileges via heartbeat context inheritance and senderIsOwner parameter manipulation. Attackers can exploit improper context validation to bypass sandbox restrictions and achieve unauthorized privilege escalation.

OpenClaw before 2026.3.28 contains an authorization bypass vulnerability in Discord text approval commands that allows non-approvers to resolve pending exec approvals. Attackers can send Discord text commands to bypass the channels.discord.execApprovals.approvers allowlist and approve pending host execution requests.

OpenClaw before 2026.3.31 contains a server-side request forgery vulnerability in the marketplace plugin download functionality that allows remote attackers to make arbitrary network requests. Attackers can exploit unguarded fetch() calls to access internal resources or interact with external services on behalf of the affected system.

OpenClaw versions 2026.3.22 before 2026.3.31 contain a signature verification bypass vulnerability in the Nostr DM ingress path that allows pairing challenges to be issued before event signature validation. An unauthenticated remote attacker can send forged direct messages to create pending pairing entries and trigger pairing-reply attempts, consuming shared pairing capacity and triggering bounded relay and logging work on the Nostr channel.

OpenClaw before 2026.3.31 contains a trust-decline vulnerability that preserves attacker-discovered endpoints in remote onboarding flows. Attackers can route gateway credentials to malicious endpoints by having their discovered URL survive the trust decline process into manual prompts requiring operator acceptance.

OpenClaw before 2026.3.28 contains an authorization bypass vulnerability in the chat.send gateway method where ACP-only provenance fields are gated by self-declared client metadata from WebSocket handshake rather than verified authorization state. Authenticated operator clients can spoof ACP identity labels and inject reserved provenance fields intended only for the ACP bridge by manipulating client metadata during connection.

OpenClaw before 2026.4.2 fails to enforce write scopes on the POST /sessions/:sessionKey/kill endpoint in identity-bearing HTTP modes. Read-scoped callers can terminate running subagent sessions by sending requests to this endpoint, bypassing authorization controls.

OpenClaw before 2026.3.31 contains a server-side request forgery vulnerability in the marketplace plugin download functionality that allows attackers to access internal resources by following unvalidated redirects. The marketplace.ts module fails to restrict redirect destinations during archive downloads, enabling remote attackers to redirect requests to arbitrary internal or external servers.

OpenClaw before 2026.3.31 contains a time-of-check-time-of-use race condition in the remote filesystem bridge readFile function that allows sandbox escape. Attackers can exploit the separate path validation and file read operations to bypass sandbox restrictions and read arbitrary files.

OpenClaw before 2026.4.2 contains an improper trust boundary vulnerability allowing untrusted workspace channel shadows to execute during built-in channel setup and login. Attackers can clone a workspace with a malicious plugin claiming a bundled channel id to achieve unintended in-process code execution before the plugin is explicitly trusted.

OpenClaw before 2026.3.28 loads the current working directory .env file before trusted state-dir configuration, allowing environment variable injection. Attackers can place a malicious .env file in a repository or workspace to override runtime configuration and security-sensitive environment settings during OpenClaw startup.

In OpenBSD through 7.8, the slaacd and rad daemons have an infinite loop when they receive a crafted ICMPv6 Neighbor Discovery (ND) option (over a local network) with length zero, because of an "nd_opt_len * 8 - 2" expression with no preceding check for whether nd_opt_len is zero.

OpenClaw before 2026.4.2 accepts non-loopback cleartext ws:// gateway endpoints and transmits stored gateway credentials over unencrypted connections. Attackers can forge discovery results or craft setup codes to redirect clients to malicious endpoints, disclosing plaintext gateway credentials.

Glances is an open-source system cross-platform monitoring tool. Prior to version 4.5.4, the Cassandra export module (`glances/exports/glances_cassandra/__init__.py`) interpolates `keyspace`, `table`, and `replication_factor` configuration values directly into CQL statements without validation. A user with write access to `glances.conf` can redirect all monitoring data to an attacker-controlled Cassandra keyspace. Version 4.5.4 contains a fix.

Glances is an open-source system cross-platform monitoring tool. Prior to version 4.5.4, a Server-Side Request Forgery (SSRF) vulnerability exists in the Glances IP plugin due to improper validation of the public_api configuration parameter. The value of public_api is used directly in outbound HTTP requests without any scheme restriction or hostname/IP validation. An attacker who can modify the Glances configuration can force the application to send requests to arbitrary internal or external endpoints. Additionally, when public_username and public_password are set, Glances automatically includes these credentials in the Authorization: Basic header, resulting in credential leakage to attacker-controlled servers. This vulnerability can be exploited to access internal network services, retrieve sensitive data from cloud metadata endpoints, and/or exfiltrate credentials via outbound HTTP requests. The issue arises because public_api is passed directly to the HTTP client (urlopen_auth) without validation, allowing unrestricted outbound connections and unintended disclosure of sensitive information. Version 4.5.4 contains a patch.