Fills in the result summary on the task from executor response.
Single pass: find section "Result summary" → task.result = section content. If no "Result summary" section, appends to issues. No parsing of filename or other details. Used by legacy/test paths; normal execution uses exec_extract(Tool) and task.result from exec_runs.
Content we expect (task_execution.md):
## Result summary
We have the information we need; it is complete.| task |
the task to set result on |