fix: strip -p/--print flags in agentic mode so Claude can actually modify files

The agentic invocation path inherited -p (print mode) from _CLAUDE_BASE_ARGS
but only stripped the stdin sentinel "-". Print mode makes Claude a one-shot
text completer that cannot use tools or write files, resulting in zero diffs.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

cross_eval/agent.py

@@ -419,8 +419,11 @@ def invoke_agent_agentic(
     if agent.reasoning_effort and _supports_reasoning_effort(agent.command):
         cmd.extend(["-c", f'model_reasoning_effort="{agent.reasoning_effort}"'])
 
-    # Strip stdin sentinel ("-") from args for agentic mode
-    args = [a for a in agent.args if a != "-"]
+    # Strip print-mode flags and stdin sentinel from args for agentic mode.
+    # -p / --print makes Claude a one-shot text completer that cannot use tools
+    # or modify files, which defeats the entire purpose of agentic execution.
+    _STRIP_FOR_AGENTIC = {"-", "-p", "--print"}
+    args = [a for a in agent.args if a not in _STRIP_FOR_AGENTIC]
     cmd.extend(args)
 
     # System prompt via flag if supported