Fix file sync integration tests and session auto-purge bug

Refactor integration tests, introduce CUJ isolation, create E2E mesh runner script with no-rebuild, and clean up orphan files

chore: remove hardcoded prod passwords from test scripts and workflows

fix: resolve file sync infinite bounce loop and unbound local variable bugs

fix: cat() on deleted file returns instantly instead of timing out ...

Root cause: two separate silent-failure bugs caused 15s stalls on file-not-found:

1. agent-node/node.py _push_file(): when a READ arrives for a missing file,
   silently returned without sending any response. The Hub journal sat waiting
   for the full 15s timeout.
   Fix: send ERROR SyncStatus back immediately with task_id so the journal
   resolves at once.

2. ai-hub/assistant.py cat(): after the mirror fast-path found no file, it
   fell through and sent a READ gRPC to the node anyway, compounding the wait.
   Fix: short-circuit with {error: File not found} when the file is absent
   from the session mirror — no node round-trip needed.

test: added Case 9 (TestSmallFileSync) — asserts that cat() on a freshly
deleted file resolves in < 3s on both nodes (guard against regression).

fix(integration_tests): correct API paths to /api/v1/ and make USER_ID env-configurable ...

- SESSIONS_PATH and NODES_PATH now include /api/v1/ prefix matching production
- USER_ID reads from SYNC_TEST_USER_ID env var (required to pass node access checks)
- All 8 file sync tests pass against live production (verified via SSH)

test: add file sync integration test suite (cases 1-8) ...

Covers all 8 real-world file sync scenarios:
  Cases 1-4 (small files):
  1. Write from node-1 → node-2 + server mirror receive it
  2. Write from server (via node-1) → all client nodes receive it
  3. Delete from server → all client nodes purged
  4. Delete from node-2 → server mirror + node-1 purged

  Cases 5-8 (20 MB large files, multi-chunk):
  5-8: Same cases as above but with 20 MB files to exercise multi-chunk sync

  Design:
  - Synchronous httpx client (no asyncio overhead)
  - Shared module-scoped swarm_control session with both nodes attached
  - _poll_until() helper for timeout-based convergence checks
  - Auto-skip via requires_nodes marker when Hub/nodes are unreachable
  - Env vars: SYNC_TEST_BASE_URL, SYNC_TEST_NODE1, SYNC_TEST_NODE2