feat(repo-cache): sync pre-fetched repositories without listing the organization

## Problem

Following #27, `repo_cache.sync_org()` correctly reports empty repositories separately from synchronization failures. However, it always lists the organization through `gh` and accepts only repository-name filters.

Consumers such as `generate-repo-overview` already fetch the complete repository set through PyGithub. They need to synchronize those same repositories, but using `sync_org()` causes a second organization listing and loses the already-fetched repository records.

## Proposed change

Add a high-level synchronization API that accepts repository records supplied by the caller, for example:

```python
sync_repositories(
*,
org: str,
repositories: Sequence[Repository],
cache_dir: Path,
workers: int = DEFAULT_SYNC_WORKERS,
progress: Callable[[str], None] | None = None,
) -> SyncReport
```

`sync_org()` should retain its current behavior and delegate to this API after listing and filtering repositories.

## Acceptance criteria

- Callers can synchronize a pre-fetched sequence of `repo_cache.Repository` records without another organization-list API request.
- The returned `SyncReport` preserves the post-#27 empty-repository and failure classification.
- Checkout paths and outcome ordering remain deterministic and compatible with `sync_org()`.
- `sync_org()` behavior remains unchanged and shares the implementation.
- The library documentation includes the pre-fetched-repository use case.

This would let downstream collectors fetch GitHub metadata once, synchronize checkouts once, and pass the resulting paths into their own processing workers.