CBST2-01: Fix JWT Replay Attack Vulnerability#297
CBST2-01: Fix JWT Replay Attack Vulnerability#297ManuelBilbao wants to merge 20 commits intosigp-audit-fixesfrom
Conversation
ltitanb
left a comment
ltitanb
left a comment
There was a problem hiding this comment.
In the current implementation, the nonce is effectively the total number of requests from a given module. This means that whenever any of server / module is restarted, everything has to be restarted, otherwise the client / server nonces would be out of sync. It also means that JWTs are regenerated for each request.
A different approach could be:
- only include the request hash in the JWT, this would allow separate restarting but not protect against replay
- use separate JWT with shorter expiration, say 30 seconds, and a separate header for replay protection with nonce + hash, with the nonce reset once the JWT is renewed
There was a problem hiding this comment.
these seem changes due to a different formatter you're using? we have a taplo.toml in the root which should take care of it
There was a problem hiding this comment.
Actually I run taplo fmt Cargo.toml and that was the output. Do you have other results? For context, my installed version is taplo 0.10.0
There was a problem hiding this comment.
Testing a bit more, the only difference between my editor's format and the taplo command is the indentation (my editor use 4 spaces versus the 2 spaces used by the command). This can be solved setting the indent_string param in the taplo.toml config with the desired value
|
|
||
| Ok(()) | ||
| fn build_jwt(&self) -> Result<Jwt, SignerClientError> { | ||
| create_jwt(&self.module_id, self.nonce, &self.jwt_secret).map_err(SignerClientError::from) |
There was a problem hiding this comment.
why not increasing the nonce here directly?
3 participants
jticlaim as a nonce to avoid reply attacks.