# Changelog

## Score

- CAI 54 → 55 (+1.0)
- Rubric changed (rubric-2026.10.5 → rubric-2026.11.4) — scores are not directly comparable.

## Lenses

- Code Health 71 → 71 (-0.1)
- Architecture 93 → 93 (-0.0)
- Maturity 43 → 45 (+2.7)
- Readiness 54 → 53 (-0.9)
- Security 69 → 69 (+0.0)

## Resolved (6)

- Ambiguous distinction between 'get' and 'getInternal'. Without documentation, it is unclear why a separate method is needed for internal claims versus standard claims, suggesting a potential leakage of implementation details or a confusing API contract.
- Duplicate method names for the same operation. Both methods accept an array of custom claims and likely perform the same assignment or storage logic.
- Inconsistent error handling strategy. `checkOrFail` implies throwing an exception on failure, while `check` likely returns a boolean or nullable result. While distinct, the naming convention `check` vs `checkOrFail` is common but `check` returning a boolean is often less intuitive than `isValid` or `tryCheck` in modern APIs.
- Inconsistent naming for retrieving the current user. `JWTAuth` uses `user()`, while `JWTGuard` provides both `user()` and `getUser()`. This creates confusion about which method to use in different contexts (Facade vs Guard instance).
- Off-boarding risk: anonymized user #1
- Redundant method for strict matching. `matchesStrict` appears to be a convenience wrapper for `matches` with `strict=true`, adding no distinct functional value while increasing API surface.

## New (2)

- Edited copy of a member (13 corresponding lines) (src/Providers/JWT/Lcobucci.php)
- Off-boarding risk: anonymized user #1
