会议议程生成助手

This skill should be used when the user provides meeting-related information (title, venue, time, agenda items, participants, etc.) and wants to generate a f...

This skill appears to implement a legitimate meeting-agenda generator, but take the following precautions before installing or running it: - Review and run the script in a sandboxed environment (or a disposable virtualenv) first. The script runs pip install programmatically and may modify your Python environment. - Note the documentation/implementation mismatch: SKILL.md mentions installing 'reportlab' while the script installs/uses 'docx2pdf'. Confirm which packages you want installed and consider installing them yourself (pip install python-docx docx2pdf) in a controlled environment before letting the script auto-install. - Inspect the remainder of the script (the file was truncated in the listing). Ensure the main() entrypoint does nothing unexpected (network calls, reading unrelated files, executing arbitrary commands). - Because the script can invoke external converters (Win Word COM, LibreOffice via subprocess), running it may execute local binaries; only run where those binaries are trusted. - If you want minimal impact, run the script inside a temporary virtual environment or container so pip installs do not affect your global Python environment. If you want, I can: - show the remainder of the script's main() body if you provide it, and re-check for unexpected behavior; - suggest an easy sandbox command set to run the script safely; or - produce a fixed SKILL.md that matches the actual implementation.

Type: OpenClaw Skill Name: meeting-agenda-tool Version: 1.0.0 The skill is a legitimate tool for generating meeting agendas in Word and PDF formats. The Python script `scripts/generate_agenda.py` uses standard libraries like `python-docx` and `docx2pdf` to create documents based on user-provided JSON data. While the script performs automated dependency installation via `pip` and executes system commands to invoke LibreOffice for PDF conversion, these actions are directly aligned with the stated purpose in `SKILL.md` and are implemented using safe practices (e.g., passing arguments as lists to `subprocess.run` to prevent shell injection). No evidence of data exfiltration, persistence, or malicious intent was found.