# Changelog

## Score

- CAI 60 → 60 (+0.2)
- Rubric changed (rubric-2026.10.5 → rubric-2026.11.2) — scores are not directly comparable.

## Lenses

- Code Health 62 → 62 (+0.0)
- Architecture 97 → 97 (+0.0)
- Maturity 51 → 52 (+1.3)
- Readiness 63 → 63 (+0.0)
- Security 79 → 74 (-4.9)

## Resolved (2)

- Ambiguous naming between configuration object access and configuration action. `SSHKit.config` is a Property (getter), while `SSHKit.configure` is a Method (setter/block). However, `SSHKit.Backend.Abstract` and `SSHKit.Backend.Netssh` expose both `config()` and `configure()` as Methods. It is unclear if `config()` on the backend returns the configuration object or triggers a configuration action, creating confusion with the top-level `SSHKit` API where `config` is a property and `configure` is a method.
- Inconsistent naming for state predicates. `started?` and `complete?` are state checks. `success?` and `failure?` are also state checks but imply a final outcome. While distinct, the presence of `started?` (boolean) and `complete?` (boolean) alongside `success?` (boolean) is consistent. However, `Command.started` is also a setter `started=(new_started)`. This is fine. The inconsistency is minor: `Command.complete?` vs `Command.success?`. If a command is complete, it is either successful or failed. Having both `complete?` and `success?`/`failure?` is redundant if `complete?` implies the command has finished, but it is useful. The real issue is `Command.started?` vs `Command.started=`. This is standard. No major inconsistency here.
