RE: [security-services] FW: <fyi> HMAC flaw in XML DSig, Redux (W3C Blog)

["Scott Cantor" <cantor.2@osu.edu>]

robert.philpott@rsa.com wrote on 2009-07-16:
> I'm not aware of any implementations that generate messages using
HMAC-based
> signatures.  However, the issue isn't so much what SAML implementations
> might send out, it's how they would respond if they received a hacked
> message that contained an HMAC-based signature.  Depending on the
> implementation's underlying DSIG package, it could conceivably accept such
a
> message as having a valid signature and attempt to process its contents.

Sure, but only if it were actually using HMAC explicitly as a RP, since it
would have to validate the signing key. Obviously the attack lets you spoof
the key, but you'd have to be expecting to validate a symmetric key to begin
with.

-- Scott