This is a comment left during a code review.
Path: internal/utils/filename_test.go
Line: 159-183

Comment:
**`debugOnce` may already be spent when this test runs**

`debugOnce` is a package-level `sync.Once` that fires exactly once for the entire test binary run. `TestCleanupLogs` (in `debug_test.go`) defers `utils.ConfigureDebug(config.GetLogsDir())`, which restores `logsDir` to the real system logs path before returning. If any subsequent test calls `Debug(...)` with that `logsDir` set (e.g., `TestDetermineFilename_PriorityOrder` calls `DetermineFilename` which calls `Debug` on several code paths), `debugOnce.Do` fires against the real logs directory — not a temp directory.

When `TestDetermineFilename_LoggingIntegration` then calls `ConfigureDebug(tempDir)` and invokes `DetermineFilename`, `debugOnce.Do` is already a no-op. All `Debug` writes still go to the original file (or to `nil` if that file couldn't be created), never to `tempDir`. The test therefore only validates the filename return value — already covered by `TestDetermineFilename_PriorityOrder` — rather than the debug-logging integration it is named after.

To reliably test the logging path, the debug subsystem needs a way to reset its state between tests (e.g., a `resetDebugForTesting()` function that zeroes out both `logsDir` and `debugOnce`), or the test should verify that a log file was actually written to `tempDir`:

```go
entries, err := os.ReadDir(tempDir)
if err != nil || len(entries) == 0 {
    t.Error("expected a debug log file to be written to the configured temp dir")
}
```

Without such an assertion, a future regression in the `Debug`→file-write path would not be caught by this test.

How can I resolve this? If you propose a fix, please make it concise.

---

This is a comment left during a code review.
Path: cmd/root_startup.go
Line: 56-60

Comment:
**Retention of `0` still produces one log after cleanup**

The guard `if retention > 0` correctly avoids underflow, but when `retention == 0` the else-branch calls `CleanupLogs(0)` (deletes all existing logs) and then a new log is created immediately after. A user who explicitly sets `log_retention_count: 0` expecting zero persistent logs will end up with one log per session.

`CleanupLogs` already interprets negative values as "keep all logs". If `0` is intended to mean "keep none / disable logging", the startup code should also skip `ConfigureDebug` entirely (or call `ConfigureDebug("")`) in that case:

```go
if retention == 0 {
    // User requested no log persistence; skip logging entirely
    utils.ConfigureDebug("")
    return nil
}
utils.ConfigureDebug(logsDir)
utils.CleanupLogs(retention - 1)
```

If `0` is instead meant as "use default", that should be documented in the settings struct and handled accordingly.

How can I resolve this? If you propose a fix, please make it concise.