Improve bd-my64 fix based on oracle review

Oracle identified a critical race condition in the initial fix:
- Pre-push hook checked for changes but didn't flush first
- Pending 5s-debounced flushes could land after the check
- Result: stale JSONL could still be pushed

Improvements:
1. Pre-push now flushes pending changes FIRST (bd sync --flush-only)
2. Uses git status --porcelain to catch ALL change types:
   - Staged, unstaged, untracked, deleted, renamed, conflicts
3. Handles both beads.jsonl and issues.jsonl (backward compat)
4. Works even without bd installed (git-only check)
5. Pre-commit stages both JSONL files (simpler loop)

This completely eliminates the race condition.

examples/git-hooks/README.md

@@ -69,18 +69,20 @@ The hook is silent on success, fast (no git operations), and safe (fails commit
 
 ### pre-push
 
-Before each push, the hook checks:
+Before each push, the hook:
 
 ```bash
-git diff --quiet .beads/beads.jsonl
+bd sync --flush-only  # Flush pending changes (if bd available)
+git status --porcelain .beads/*.jsonl  # Check for uncommitted changes
 ```
 
 This prevents pushing stale JSONL by:
-1. Checking if JSONL has uncommitted changes (working tree or staging)
-2. Failing the push with clear error message if changes exist
-3. Instructing user to commit JSONL before pushing again
+1. Flushing pending in-memory changes from daemon's 5s debounce
+2. Checking for uncommitted changes (staged, unstaged, untracked, deleted)
+3. Failing the push with clear error message if changes exist
+4. Instructing user to commit JSONL before pushing again
 
-This solves bd-my64: changes made between commit and push are caught before reaching remote.
+This solves bd-my64: changes made between commit and push (or pending debounced flushes) are caught before reaching remote.
 
 ### post-merge