OpenClaw versions prior to 2026.5.4 contain a vulnerability in the bundled device-pair plugin that allows authorized non-owner chat senders to issue device-pairing bootstrap codes. This enables these users to enroll devices with operator/node capabilities, which retain persistent credentials until removed. The vulnerability requires the attacker to have chat command access and does not affect unauthenticated users. The issue is fixed in version 2026.5.4.
A vulnerability in OpenClaw's scoped Gateway chat.send route allows commands requiring higher privileges (operator.approvals or operator.admin) to be executed with only operator.write scope when routed through inherited external delivery routes. This affects deployments where scoped Gateway clients with operator.write can deliver commands into sessions with external routes. The issue is fixed in version 2026.5.18. Until upgraded, avoid granting operator.write tokens to clients that can deliver commands into sessions with external routes unless fully trusted.
An attacker who can send HTML chat messages (via Matrix or XMPP) can inject arbitrary styled content, phishing links, and CSS that manipulates the chat UI. This vulnerability was fixed in Thunderbird 152.0.1 and Thunderbird 140.12.1.
The PyPI package 'pdf-converter-pro' version 1.0.0 is a malicious package that masquerades as a PDF conversion utility but contains no actual PDF generation code. Instead, it exfiltrates Python source code files from the user's home directory, current directory, and filesystem roots by sending them to an attacker-controlled Telegram chat using a hardcoded bot token. The package tracks already exfiltrated files locally to avoid duplicates and does not provide the advertised functionality. This behavior is triggered when the package's main method is called with specific literal arguments.
Rocket.Chat is an open-source, secure, fully customizable communications platform. Prior to 8.5.0, 8.4.1, 8.3.3, 8.2.3, 8.1.4, 8.0.5, 7.13.7, and 7.10.11, an unauthenticated network attacker obtains a valid Rocket.Chat OAuth access token for an arbitrary user by sending a single HTTP POST with MongoDB query operators to /oauth/token. The Rocket.Chat OAuth2 server does not validate that grant parameters are strings before forwarding them to findOne({...}) against the oauth_apps and oauth_access_tokens collections, so an attacker substitutes {"$ne": null} for client_id, client_secret, and refresh_token and receives a freshly minted {access_token, refresh_token} pair bound to whichever user's refresh token Mongo returned first. The resulting access token is a first-class bearer credential against the full /api/v1/* surface as that user. By iterating with $nin / $regex operators the attacker walks the entire oauth_access_tokens collection, collecting one fresh access token per user per request. If any matched token belongs to an admin, the stolen bearer gives full admin API access (including Apps-Engine app installation, i.e. server-side code execution). No account, credentials, userId, or prior interaction with the instance are required. This vulnerability is fixed in 8.5.0, 8.4.1, 8.3.3, 8.2.3, 8.1.4, 8.0.5, 7.13.7, and 7.10.11.
Rocket.Chat is an open-source, secure, fully customizable communications platform. Prior to 8.5.0, 8.4.1, 8.3.3, 8.2.3, 8.1.4, 8.0.5, 7.13.7, and 7.10.11, Rocket.Chat's sendFileMessage DDP method passes the entire attacker-supplied file object into Uploads.updateFileComplete, which merges it directly into a MongoDB $set update via Object.assign. There is no allow-list of writable fields. An attacker can therefore rewrite any column on their own upload record, notably store and the store-specific path fields. This vulnerability is fixed in 8.5.0, 8.4.1, 8.3.3, 8.2.3, 8.1.4, 8.0.5, 7.13.7, and 7.10.11.
Line Desktop MCP is a project that, while unaffiliated with the official line-bot-mcp-server, allows users to directly operate the LINE Desktop application on Windows or Mac via MCP. `line-desktop-mcp` supports a `--http-mode` Streamable HTTP transport for use with clients such as n8n. In this mode the server binds to `0.0.0.0` and exposes the MCP `/mcp` endpoint without an MCP-layer authentication check. Prior to version 1.1.2, any network client that can reach the port can initialize a session, list tools, and call tools that read LINE Desktop chat history or send LINE messages through the already logged-in desktop application. Version 1.1.2 fixes the issue.
---
_-= Per source details. Do not edit below this line.=-_
## Source: amazon-inspector (d5ade836e7f92049242a01dbc0782900900c4e28eb7e08f9d9ebc611aab80762)
nodepathbalance54 exports a single function (`nodeaxionweb`) whose implementation is hidden inside a hand-rolled stack-based JavaScript VM in index.js (functions iL/iw with ~200 opcodes, base64-encoded bytecode chunks in array S). Decoding the VM's constant pool reveals a hardcoded Telegram bot token (8604230339:AAE0jcxBc1QC_YV1Um6TT4bvC0mgZGlhoy4), chat id (-1002861294159), the base URL https://api.telegram.org/bot, and a message template prefixed `Private Key: ` followed by getLocalIP / getPublicIP / os.hostname / platform / arch / ISO timestamp fields, posted via axios.post to /sendMessage. Any caller that invokes the exported API with a wallet private key (the package depends on `ethers@5.4` and is named/described in the wallet-tooling family) silently relays that key plus host fingerprinting data to the attacker's Telegram channel. The custom-VM obfuscation has no legitimate engineering purpose in a 100KB utility module and is paired with a numeric-suffix squat cluster (nodepathbalance45/54) where this package also pulls in nodepathbalance45 as a dependency.
## Source: ghsa-malware (cdc355f0a37fdea9334e29b85d8ee4fa970dfc51471cd95d9ebaa5d1f7e7e9ee)
Any computer that has this package installed or running should be considered fully compromised. All secrets and keys stored on that computer should be rotated immediately from a different computer. The package should be removed, but as full control of the computer may have been given to an outside entity, there is no guarantee that removing the package will remove all malicious software resulting from installing it.
Lumiverse is a full-featured AI chat application. Prior to 0.9.7, the MCP server creation endpoint validates the command field against an allowlist of binary names but forwards the args array to the child process without any validation. Every binary on the allowlist accepts an inline-code execution flag (-e for node/bun, -c for python3/deno), giving any logged-in user arbitrary OS-level code execution on the Lumiverse server. The route requires only requireAuth (not requireOwner). The server binds on all interfaces (::) and the host-header rebinding check is bypassed trivially by any HTTP client that sends Host: localhost:<port> directly, making this exploitable from any machine with network access to the server port. This vulnerability is fixed in 0.9.7.
The Essential Chat Support plugin for WordPress is vulnerable to authorization bypass in all versions up to, and including, 1.0.1. This is due to the plugin not properly verifying that a user is authorized to perform an action. This makes it possible for unauthenticated attackers to reset all plugin configuration settings — including general settings, display rules, custom CSS, and WooCommerce tab settings — to their defaults by sending a POST request with ecs_reset_settings=1.
New API is a large language mode (LLM) gateway and artificial intelligence (AI) asset management system. In versions 0.11.9-alpha.1 and prior, the SSRF protection introduced in v0.9.0.5 (CVE-2025-59146) and hardened in v0.9.6 (CVE-2025-62155) does not block the unspecified address 0.0.0.0. A regular (non-admin) user holding any valid API token can send a multimodal request to /v1/chat/completions, /v1/responses, or /v1/messages with 0.0.0.0 as the image/file URL host, bypassing the private-IP filter and causing the server to issue HTTP requests to localhost. This constitutes at minimum a blind SSRF; when the request is routed through an AWS/Bedrock Claude adaptor, the fetched content is inlined into the model response, upgrading it to a full-read SSRF. At time of publication, there are no publicly available patches.
### Impact
A database query used during proxy API key checks mixed the caller-supplied key value into the query text instead of passing it as a separate parameter. An unauthenticated attacker could send a specially crafted `Authorization` header to any LLM API route (for example `POST /chat/completions`) and reach this query through the proxy's error-handling path.
An attacker could read data from the proxy's database and may be able to modify it, leading to unauthorised access to the proxy and the credentials it manages.
### Patches
Fixed in **`1.83.7`**. The caller-supplied value is now always passed to the database as a separate parameter. Upgrade to `1.83.7` or later.
### Workarounds
If upgrading is not immediately possible, set `disable_error_logs: true` under `general_settings`. This removes the path through which unauthenticated input reaches the vulnerable query.
### References
- Patched release: [`v1.83.7-stable`](https://github.com/BerriAI/litellm/releases/tag/v1.83.7-stable)
**Discovery Credit**: Tencent YunDing Security Lab
CVE-2026-8125 is a medium-severity SQL injection vulnerability in code-projects Simple Chat System version 1.0. It affects the sendMessage.php file, where manipulation of certain parameters can lead to SQL injection. The vulnerability can be exploited remotely without user interaction and requires low privileges. Although a public exploit exists, there is no confirmed patch or official remediation available at this time.
Malicious npm package published by user `shetty123` as part of a Telegram account hijacking framework targeting Indian Telegram users. All 502 published versions (1.0.1 through 1.3.207) are malicious. Pairs with `ams-ssk`, which provides the operator's server-side AMS/CMS infrastructure.
`common-tg-service` performs full Telegram account takeover at runtime when the service is initialized (no install-time hooks, which lets it bypass scanners that gate on preinstall/postinstall lifecycle scripts). Behavior includes: implanting a hardcoded 2FA password (`Ajtdmwajt1@`) and recovery email on hijacked accounts; polling an operator-controlled Gmail inbox over IMAP (`imap.gmail.com`) to auto-submit 2FA confirmation codes; revoking all device authorizations except the attacker's session; harvesting OTP codes by monitoring Telegram chat 777000 and forwarding them to the operator; running SRP ownership checks against managed accounts and flagging rotated 2FA as unrecoverable; and fetching remote JSON configuration from `npoint.io` so operators can change behavior without re-publishing.
Blocked outbound requests are laundered through a relay at `helper-thge.onrender.com`. Stolen accounts and updates are exfiltrated to attacker-controlled Telegram channels (`-1001801844217` and `-1001972065816`). Operator infrastructure includes `paidgirl.site`, `cms.paidgirl.site`, `report-upi.netlify.app`, and `promoteClients2.glitch.me`.
---
_-= Per source details. Do not edit below this line.=-_
## Source: amazon-inspector (e769b4a4ea131df52366622b66d63dfbc7c1f5642ac8478c37f86d6c25477bb8)
The package presents itself as a generic 'Common Telegram service for NestJS applications' but functions as one party's production application published as a library. Multiple installer-harming behaviors are present:
1. Authentication backdoor. AppModule registers AuthGuard as a global APP_GUARD. dist/guards/auth.guard.js line 84 accepts the literal string 'santoor' as a valid x-api-key (or apiKey query parameter) and grants access to any route. Any NestJS service that imports AppModule from this package exposes every endpoint to anyone who sends `x-api-key: santoor`. The same value is baked into the shipped Swagger UI as the default auth value.
2. Silent relay of consumer outbound HTTP. dist/utils/fetchWithTimeout.js (used package-wide) automatically POSTs the original request (URL, method, headers, body, params) to https://helper-thge.onrender.com/execute-request whenever an upstream call returns 403 or 495, unless the consumer overrides process.env.bypassURL. The relay endpoint is author-controlled and authenticated with the same 'santoor' shared secret, so the author can read every mirrored request — including any auth headers or bodies the consumer's service was sending to its own upstreams.
3. Default proxy and IP-management routing through author infrastructure. dist/components/Telegram/utils/generateTGConfig.js line 97 defaults PROXY_API_URL to https://cms.paidgirls.site/ip-management and PROXY_API_KEY to 'santoor'. With proxying enabled, all MTProto traffic for the consumer's Telegram sessions is routed through SOCKS5 proxies chosen by the author, who can intercept, drop, or correlate the consumer's Telegram identities.
4. Remote zip drop into the consumer's working directory. dist/cloudinary.js line 87 fetches https://cms.paidgirls.site/folders/<folderName>/files/download-all, writes the response to process.cwd()/temp.zip, and calls zip.extractAllTo(process.cwd(), true) — overwriting any files in the consumer's working directory with whatever the author's server returns. There is no hash check, signature, or version pinning, and the destination is mutable. If extracted files land where the consumer's process later loads code or config, this is full RCE in the consumer's environment.
5. Hardcoded couplings that are not disclosed. dist/guards/auth.guard.js lines 15-26 hardcode ALLOWED_IPS and ALLOWED_ORIGINS for paidgirls.site, zomcall.netlify.app, and tgchats.netlify.app, and the package wires logging notifications to specific Telegram channels owned by the author.
Taken together, importing AppModule from this package gives the author master authentication on the consumer's API, a persistent traffic-mirroring channel, control of the consumer's Telegram proxy path, and an ability to write arbitrary files into the consumer's process CWD on demand.
## Source: ghsa-malware (3e16628ad8dc1e6a6c98044a97ae5b72aec56f8a5a5bbc5bdff50bdaf51d978e)
Any computer that has this package installed or running should be considered fully compromised. All secrets and keys stored on that computer should be rotated immediately from a different computer. The package should be removed, but as full control of the computer may have been given to an outside entity, there is no guarantee that removing the package will remove all malicious software resulting from installing it.
OpenClaw versions before 2026.3.28 have a privilege escalation vulnerability (CWE-863) that allows authenticated operators with write permissions to access and modify admin-level Talk Voice configuration settings via the chat.send endpoint. This flaw enables operators to perform actions reserved for administrators.
CVE-2026-7090 is a medium severity cross-site scripting (XSS) vulnerability in code-projects Chat System version 1.0. It affects the /admin/send_message.php file where the 'msg' parameter can be manipulated to inject malicious scripts. The vulnerability can be exploited remotely and requires user interaction with elevated privileges. No official patch or remediation guidance is currently available, and no known exploits are reported in the wild.
OpenClaw versions before 2026.4.2 have a vulnerability in the Zalo webhook replay deduplication mechanism. This flaw allows legitimate events from different conversations or senders to be incorrectly treated as duplicates, causing silent message suppression and disruption of bot workflows across chat sessions.
OpenClaw versions before 2026.3.28 have a privilege escalation vulnerability in the chat.send endpoint. This flaw allows users with write-scoped gateway access to persist admin-only verboseLevel session overrides via the /verbose parameter. Exploiting this can bypass access controls and expose sensitive information intended only for administrators.
OpenClaw versions before 2026.3.28 have an authorization bypass vulnerability in the chat.send gateway method. The flaw arises because ACP-only provenance fields are controlled by client metadata from the WebSocket handshake, which is self-declared and not verified against the actual authorization state. This allows authenticated operator clients to spoof ACP identity labels and inject reserved provenance fields intended only for the ACP bridge by manipulating client metadata during connection.