Remove dead code: execute_level, render_dog_from_cat, duplicate file_hash

- Remove execute_level() from tasks/execute.py (defined but never called)
- Remove render_dog_from_cat() from legacy_tasks.py (test convenience, never used)
- Remove duplicate file_hash() from legacy_tasks.py, import from cache_manager
- Remove unused hashlib import

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

tasks/execute.py

@@ -379,51 +379,3 @@ def execute_step(
         }
 
 
-@app.task(bind=True, name='tasks.execute_level')
-def execute_level(
-    self,
-    steps_json: List[str],
-    plan_id: str,
-    cache_ids: Dict[str, str],
-) -> dict:
-    """
-    Execute all steps at a given dependency level.
-
-    Steps at the same level can run in parallel.
-
-    Args:
-        steps_json: List of JSON-serialized ExecutionSteps
-        plan_id: ID of the parent execution plan
-        cache_ids: Mapping from step_id to cache_id
-
-    Returns:
-        Dict with results for all steps
-    """
-    from celery import group
-
-    # Dispatch all steps in parallel
-    tasks = [
-        execute_step.s(step_json, plan_id, cache_ids)
-        for step_json in steps_json
-    ]
-
-    # Execute in parallel and collect results
-    job = group(tasks)
-    results = job.apply_async()
-
-    # Wait for completion
-    step_results = results.get(timeout=3600)  # 1 hour timeout
-
-    # Build cache_ids from results
-    new_cache_ids = dict(cache_ids)
-    for result in step_results:
-        step_id = result.get("step_id")
-        cache_id = result.get("cache_id")
-        if step_id and cache_id:
-            new_cache_ids[step_id] = cache_id
-
-    return {
-        "status": "completed",
-        "results": step_results,
-        "cache_ids": new_cache_ids,
-    }