Describe the bug

1. content in dagu/base.yaml:
   env:
   - QUANT_SIGNAL_DIR: "/work/quant-signal"

2. content in dags/signal.yaml:
   working_dir: ${QUANT_SIGNAL_DIR}
   steps:
   - id: begin_step
   command: echo "begin_step"
   - id: run_signals
   command: ${PYTHON_BIN} ${PROJECT_DIR}/signals/run_signals.py

3. content in /work/quant-signal/.env:
   PYTHON_BIN=/usr/local/bin/python
   PROJECT_DIR=/work/quant-signal

4. The signal workflow failed at the run_signals step due to certain issue. I don't want to rerun it from the beginning after the issue is fixed. I clicked the "Retry from this step" button on the run_signals step. But it failed again with the following log: "exit status 127 recent stderr (tail): zsh:1: no such file or directory: /signals/run_signals.py", which means ${PYTHON_BIN} ${PROJECT_DIR} environment variables were not loaded when retry from the failed step.

Environment

• Dagu version: dagu_2.7.3_linux_arm64
• OS: Debian 12.5
• Installation method: download binary

Screenshots