Bon, sachant que sebsauvage lit rarement ses mails, je vais tenter de le contacter par ici. :)

Ton implémentation est non sécurisée. C'est mal car d'autres personnes risquent de s'en inspirer sans savoir les problèmes liés à cette implémentation.

1) " Warning - When comparing the output of hexdigest() to an externally-supplied digest during a verification routine, it is recommended to use the compare_digest() function instead of the == operator to reduce the vulnerability to timing attacks." https://docs.python.org/3.4/library/hmac.html#hmac.HMAC.hexdigest

2) "The digestmod argument to the hmac.new() function may now be any hash digest name recognized by hashlib. In addition, the current behavior in which the value of digestmod defaults to MD5 is deprecated: in a future version of Python there will be no default value. (Contributed by Christian Heimes in issue 17276.)" https://docs.python.org/3/whatsnew/3.4.html#hmac
Donc il faudrait mieux commencer à spécifier une valeur pour digestmod.

Comme suggéré par l'issue (1) et cette réponse sur SO (2), je proposerais d'utiliser SHA-256 ou SHA-512 comme suit :

hmac.compare_digest(hmac.new(key, name, digestmod=hashlib.sha256).hexdigest(), signature)

(+ remplacer partout où il faut bien sure et ne pas oublier import hashlib)

(1) https://bugs.python.org/issue17276
"As of now the hash algorithm for HMAC defaults to MD5. However MD5 is considered broken. HMAC-MD5 is still ok but shall not be used in new code. Applications should slowly migrate away from HMAC-MD5 and use a more modern algorithm like HMAC-SHA256.

Therefore I propose that default digestmod should be deprecated in Python 3.4 and removed in 3.5. Starting with Python 3.5 developer are forced to choose a hash algorithm like SHA256. Our documentation shall suggest it, too."

(2) http://crypto.stackexchange.com/a/9340/18518
"Yes, there are currently no known attacks on HMAC-MD5.

[…]

However, this does not mean you should use HMAC-MD5 in new cryptosystem designs. To paraphrase Bruce Schneier, "attacks only get better, never worse." We already have practical collision attacks for MD5, showing that it does not meet its original security goals; it's possible that, any day now, someone might figure out a way to turn those into a preimage attack, which would compromise the security of HMAC-MD5. A much better choice would be to use HMAC with a hash function having no known attacks, such as SHA-2 or SHA-3."
— Permalink