The htcpHandleTstRequest function in htcp.c in Squid 2.x before 2.6.STABLE24 and 2.7 before 2.7.STABLE8, and htcp.cc in 3.0 before 3.0.STABLE24, allows remote attackers to cause a denial of service (NULL pointer dereference and daemon crash) via crafted packets to the HTCP port.
lib/rfc1035.c in Squid 2.x, 3.0 through 3.0.STABLE22, and 3.1 through 3.1.0.15 allows remote attackers to cause a denial of service (assertion failure) via a crafted DNS packet that only contains a header.
An issue was discovered in http/ContentLengthInterpreter.cc in Squid before 4.12 and 5.x before 5.0.3. A Request Smuggling and Poisoning attack can succeed against the HTTP cache. The client sends an HTTP request with a Content-Length header containing "+\ "-" or an uncommon shell whitespace character prefix to the length field-value.
An issue was discovered in Squid 2.x, 3.x, and 4.x through 4.8. Due to incorrect data management, it is vulnerable to information disclosure when processing HTTP Digest Authentication. Nonce tokens contain the raw byte value of a pointer that sits within heap memory allocation. This information reduces ASLR protections and may aid attackers isolating memory areas to target for remote code execution attacks.
An issue was discovered in Squid 3.x and 4.x through 4.8 when the appenddomain setting is used (because the appended characters do not properly interact with hostname length restrictions). Due to incorrect message processing, it can inappropriately redirect traffic to origins it should not be delivered to.
An issue was discovered in Squid 2.x through 2.7.STABLE9, 3.x through 3.5.28, and 4.x through 4.7. When Squid is configured to use Basic Authentication, the Proxy-Authorization header is parsed via uudecode. uudecode determines how many bytes will be decoded by iterating over the input and checking its table. The length is then used to start decoding the string. There are no checks to ensure that the length it calculates isn't greater than the input buffer. This leads to adjacent memory being decoded as well. An attacker would not be able to retrieve the decoded data unless the Squid maintainer had configured the display of usernames on error pages.