CVE Vulnerability Catalog

Previously, a trusted template author could write a valid template where the use of the 'yield' keyword was not correctly escaped. Now valid keyword uses are escaped and non-keyword uses are not escaped.

When a JavaScript template literal contains consecutive expressions, context tracking state was not properly reset upon entering a new expression. Now template-literal expression entries correctly reset context variables so all subsequent regular expression literals are accurately recognized and escaped.

SumatraPDF 3.6.1 and earlier has a vulnerability in ChmFile::GetCharZ() where file-controlled unsigned offsets are narrowed to signed integers without a lower-bound check. This can cause reads before the /#STRINGS buffer and deterministic application termination.

SumatraPDF 3.6.1 and earlier, and pre-release 3.7.0.20369 when WebView2 is absent, has a vulnerability in ParseProtoUrl() where the signed host component of an its:// URL is used as an index into gHtmlWindows. This can cause out-of-bounds pointer read and invalid callback dereference, terminating the process.

SumatraPDF 3.6.1 and earlier has a vulnerability in MaybeFlipBitmap() where GDI+ reads the attacker-controlled PropertyTagOrientation size into a fixed 64-byte stack buffer. Opening a crafted TIFF can overwrite stack control data, though tested builds terminate via stack cookie check.

SumatraPDF 3.6.1 and earlier has a vulnerability in InitializePolicies() where the sumatrapdfrestrict.ini path starts with gPolicyRestrictions set to Perm::All and only ORs permission bits, so the INI file never revokes permissions. Deploying with this INI file, including malformed or zero-valued settings, can silently bypass configured disk, network, printing, registry, clipboard, preference, and fullscreen restrictions.

SumatraPDF 3.6.1 and earlier has a vulnerability in FrameOnCommand() where CmdExec passes a null current-tab pointer to RunWithExe(), which dereferences WindowTab::filePath. A local process in the same interactive Windows session, at an integrity level greater than or equal to SumatraPDF's, can dispatch CmdExec over DDE or WM_COPYDATA while no document tab is open, causing abrupt process termination and loss of unsaved state.

SumatraPDF 3.7.0.22298 has four independently reachable range-validation variants in src/LitDoc.cpp that allow file-controlled offsets and sizes to overflow, narrow to negative values, or wrap before incomplete bounds checks. Opening a crafted LIT file can cause invalid pointer reads and deterministic application termination.

SumatraPDF 3.7.0.22298 has a vulnerability in LitParseHeader() in src/LitDoc.cpp where the attacker-controlled hdrLen + nPieces * 16 section offset is computed using signed 32-bit arithmetic without validating the complete result. When the aggregate calculation overflows to a negative value, pointer construction reaches an invalid read in LitU32(), causing deterministic application termination.

SumatraPDF 3.7.0.22298 has a vulnerability in src/MobiDoc.cpp where the untrusted unsigned mobiHdr.hdrLen field is narrowed to a signed integer for validation; values above INT_MAX become negative and bypass the upper-bound check. When the EXTH flag is set, the original unsigned value is reused as a pointer offset, causing DecodeExthHeader() to read beyond the record buffer. Opening a crafted MOBI file can reliably terminate the application with a native access violation.

A heap-based buffer overflow was found in Eye of GNOME (eog) in the PNG metadata reader due to improper state handling when parsing split metadata chunks. A remote attacker could entice a user into opening a specially crafted PNG file, potentially leading to arbitrary code execution or a Denial of Service (DoS) via application crash.

IBM Guardium Data Protection 12.0, 12.1, and 12.2 could allow a remote attacker to obtain sensitive information due to the use of hard-coded credentials.

IBM Guardium Data Protection 12.2.2 could allow a remote attacker to conduct a man-in-the-middle attack due to improper certificate validation.

Jivejdon through 5.0 contains a cross-site request forgery vulnerability that allows remote attackers to perform state-changing actions by abusing GET endpoints lacking anti-CSRF tokens. Attackers can lure authenticated users to crafted links targeting /account/protected/delAll, /account/protected/sub/delSub, or /message/updateAction to delete private messages and subscriptions or rename threads.

Jivejdon from commit e0306088 through commit ee67a65e lacks rate limiting on the unauthenticated /account/smsVRAction endpoint handled by SmsQQAction, allowing unlimited SMS sending. Attackers can load newAccount.jsp to set session attributes, then repeatedly call the endpoint to harass arbitrary phone numbers and exhaust the operator's Tencent Cloud SMS balance.

Jivejdon through 5.0 contains a weak password storage vulnerability that stores account passwords as unsalted MD5 digests via ToolsUtil.hash() in AccountDaoSql. Attackers who obtain the user table through database access or SQL injection can crack passwords with precomputed tables or GPU attacks.

Jivejdon through 5.0 contains an authentication bypass vulnerability that allows unauthenticated attackers to access Weibo-created accounts by deriving predictable credentials from public Weibo user IDs. OAuthAccountServiceImp.transferSina() sets the password to the first four digits of the Weibo ID, letting attackers log in through normal form login to read or post as victims.

Jivejdon through 5.0 contains a stored cross-site scripting vulnerability that allows authenticated attackers to execute JavaScript by uploading attachments with an attacker-supplied Content-Type. Attackers can upload a file declared as text/html, which UploadShowAction serves inline, and share its link to run JavaScript on the application's origin for viewing users.

Jivejdon through 5.0 contains a stored cross-site scripting vulnerability that allows authenticated attackers to inject script into private short messages because receiveshortmessage.jsp renders unfiltered message bodies. Attackers can send a short message containing script, which ToolsUtil.convertURL() passes through unchanged, to execute code in the recipient's browser when opened.

Jivejdon through 5.0 contains a stored cross-site scripting vulnerability that allows authenticated attackers to inject script by posting unsanitized forum message bodies. Message bodies are rendered by messageListBody.jsp with filter="false" and non-escaping default filters, executing script in the browser of every user viewing the thread.