list_sessions_from_store exposes invalid session IDs returned by adapters

### Summary

Store-backed session listing does not validate session IDs, unlike filesystem listing and individual store-backed lookup.

### Affected code

- `src/claude_agent_sdk/_internal/sessions.py:1525-1575`
- `src/claude_agent_sdk/_internal/sessions.py:1637-1685`
- `src/claude_agent_sdk/_internal/sessions.py:1723-1732`
- `src/claude_agent_sdk/_internal/session_summary.py:219-222`

### Current behavior

The slow path passes every `session_id` returned by `SessionStore.list_sessions()` into loads and then into `SDKSessionInfo`. The fast path similarly copies IDs from session summaries or listings into slots without `_validate_uuid()`.

By contrast:

- filesystem listing skips filenames that are not UUIDs;
- `get_session_info_from_store()` returns `None` for an invalid UUID;
- mutation helpers reject invalid IDs.

### Why this matters

A malformed, stale, or externally populated backend row becomes a public `SDKSessionInfo` and consumes an offset/limit position. The SDK can therefore return session IDs that its own get, resume, fork, and mutation APIs reject.

It also causes inconsistent behavior between otherwise equivalent disk, slow-store, and summary-fast listing paths.

### Expected behavior

All public session listing implementations should filter invalid session IDs before loading, sorting, or pagination.

### Possible fix

Validate IDs as soon as adapter results are received:

- filter `list_sessions()` entries before scheduling loads;
- filter summary entries before conversion;
- ensure malformed entries do not consume page positions.

Add conformance tests for invalid UUIDs in both optional listing methods and verify parity between fast and slow paths.

### Environment

- Repository revision: current `main` audit at SDK version 0.2.128
- Bundled CLI version: 2.1.220
- Python test suite: 1,291 passed, 5 skipped
- Ruff and mypy: clean

I searched the existing issues and pull requests using the affected symbols and behavior before filing this.

Read the affected listing paths in src/claude_agent_sdk/_internal/sessions.py and the summary conversion in src/claude_agent_sdk/_internal/session_summary.py, following _validate_uuid() and SessionStore.list_sessions(). Add conformance tests covering invalid UUIDs in both optional listing methods, then verify fast and slow paths filter malformed entries before loading, sorting, or pagination.

索引モデルが issue の本文から書いたものです。