MIT Kerberos 5 (aka krb5) 1.3.x, 1.4.x, 1.5.x, 1.6.x, 1.7.x, and 1.8.x through 1.8.3 does not properly determine the acceptability of checksums, which might allow remote attackers to modify user-visible prompt text, modify a response to a Key Distribution Center (KDC), or forge a KRB-SAFE message via certain checksums that (1) are unkeyed or (2) use RC4 keys.

The kg_accept_krb5 function in krb5/accept_sec_context.c in the GSS-API library in MIT Kerberos 5 (aka krb5) through 1.7.1 and 1.8 before 1.8.2, as used in kadmind and other applications, does not properly check for invalid GSS-API tokens, which allows remote authenticated users to cause a denial of service (NULL pointer dereference and daemon crash) via an AP-REQ message in which the authenticator's checksum field is missing.

Multiple integer underflows in the (1) AES and (2) RC4 decryption functionality in the crypto library in MIT Kerberos 5 (aka krb5) 1.3 through 1.6.3, and 1.7 before 1.7.1, allow remote attackers to cause a denial of service (daemon crash) or possibly execute arbitrary code by providing ciphertext with a length that is too short to be valid.

The asn1_decode_generaltime function in lib/krb5/asn.1/asn1_decode.c in the ASN.1 GeneralizedTime decoder in MIT Kerberos 5 (aka krb5) before 1.6.4 allows remote attackers to cause a denial of service (daemon crash) or possibly execute arbitrary code via vectors involving an invalid DER encoding that triggers a free of an uninitialized pointer.

KDC in MIT Kerberos 5 (krb5kdc) does not set a global variable for some krb4 message types, which allows remote attackers to cause a denial of service (crash) and possibly execute arbitrary code via crafted messages that trigger a NULL pointer dereference or double-free.

The Kerberos 4 support in KDC in MIT Kerberos 5 (krb5kdc) does not properly clear the unused portion of a buffer when generating an error message, which might allow remote attackers to obtain sensitive information, aka "Uninitialized stack values."

Buffer overflow in the RPC library used by libgssrpc and kadmind in MIT Kerberos 5 (krb5) 1.4 through 1.6.3 allows remote attackers to execute arbitrary code by triggering a large number of open file descriptors.

Use-after-free vulnerability in the gss_indicate_mechs function in lib/gssapi/mechglue/g_initialize.c in MIT Kerberos 5 (krb5) has unknown impact and attack vectors. NOTE: this might be the result of a typo in the source code.

Double free vulnerability in the gss_krb5int_make_seal_token_v3 function in lib/gssapi/krb5/k5sealv3.c in MIT Kerberos 5 (krb5) has unknown impact and attack vectors.

The original patch for CVE-2007-3999 in svc_auth_gss.c in the RPCSEC_GSS RPC library in MIT Kerberos 5 (krb5) 1.4 through 1.6.2, as used by the Kerberos administration daemon (kadmind) and other applications that use krb5, does not correctly check the buffer length in some environments and architectures, which might allow remote attackers to conduct a buffer overflow attack.

Stack-based buffer overflow in the svcauth_gss_validate function in lib/rpc/svc_auth_gss.c in the RPCSEC_GSS RPC library (librpcsecgss) in MIT Kerberos 5 (krb5) 1.4 through 1.6.2, as used by the Kerberos administration daemon (kadmind) and some third-party applications that use krb5, allows remote attackers to cause a denial of service (daemon crash) and probably execute arbitrary code via a long string in an RPC message.

The gssrpc__svcauth_gssapi function in the RPC library in MIT Kerberos 5 (krb5) 1.6.1 and earlier might allow remote attackers to execute arbitrary code via a zero-length RPC credential, which causes kadmind to free an uninitialized pointer during cleanup.

Integer signedness error in the gssrpc__svcauth_unix function in svc_auth_unix.c in the RPC library in MIT Kerberos 5 (krb5) 1.6.1 and earlier might allow remote attackers to execute arbitrary code via a negative length value.

Stack-based buffer overflow in the rename_principal_2_svc function in kadmind for MIT Kerberos 1.5.3, 1.6.1, and other versions allows remote authenticated users to execute arbitrary code via a crafted request to rename a principal.

The telnet daemon (telnetd) in MIT krb5 before 1.6.1 allows remote attackers to bypass authentication and gain system access via a username beginning with a '-' character, a similar issue to CVE-2007-0882.

Stack-based buffer overflow in the krb5_klog_syslog function in the kadm5 library, as used by the Kerberos administration daemon (kadmind) and Key Distribution Center (KDC), in MIT krb5 before 1.6.1 allows remote authenticated users to execute arbitrary code and modify the Kerberos key database via crafted arguments, possibly involving certain format string specifiers.

Double free vulnerability in the GSS-API library (lib/gssapi/krb5/k5unseal.c), as used by the Kerberos administration daemon (kadmind) in MIT krb5 before 1.6.1, when used with the authentication method provided by the RPCSEC_GSS RPC library, allows remote authenticated users to execute arbitrary code and modify the Kerberos key database via a message with an "an invalid direction encoding".

The RPC library in Kerberos 5 1.4 through 1.4.4, and 1.5 through 1.5.1, as used in Kerberos administration daemon (kadmind) and other products that use this library, calls an uninitialized function pointer in freed memory, which allows remote attackers to cause a denial of service (crash) and possibly execute arbitrary code via unspecified vectors.

The (1) krshd and (2) v4rcp applications in (a) MIT Kerberos 5 (krb5) up to 1.5, and 1.4.x before 1.4.4, when running on Linux and AIX, and (b) Heimdal 0.7.2 and earlier, do not check return codes for setuid calls, which allows local users to gain privileges by causing setuid to fail to drop privileges using attacks such as resource exhaustion.

The (1) ftpd and (2) ksu programs in (a) MIT Kerberos 5 (krb5) up to 1.5, and 1.4.x before 1.4.4, and (b) Heimdal 0.7.2 and earlier, do not check return codes for setuid calls, which might allow local users to gain privileges by causing setuid to fail to drop privileges. NOTE: as of 20060808, it is not known whether an exploitable attack scenario exists for these issues.