Multi-panel Figure Assembler

Assemble 6 sub-figures (A–F) into a high-resolution composite figure with consistent labels, padding, and publication-ready DPI.

This skill appears coherent and implements exactly what it claims: assembling six image panels into a composite using local Python code. Before running, consider these practical precautions: (1) Review the included scripts (scripts/main.py and scripts/example.py) yourself — they will execute on your machine. (2) Run in an isolated environment (virtualenv or container) and install dependencies (pip install Pillow numpy); note requirements.txt contains a redundant/incorrect 'pil' entry — install 'Pillow' instead. (3) Use python -m py_compile scripts/main.py and run the --help to confirm expected behavior; the SKILL.md already suggests these checks. (4) Verify font fallbacks and that input paths are correct; the code already rejects ../ traversal. (5) Do not run as root and avoid executing third-party code on sensitive hosts without review. If you want higher assurance, run the example outputs in a sandbox and inspect the saved images before using in production.

Type: OpenClaw Skill Name: multi-panel-figure-assembler Version: 1.0.0 The skill is a legitimate utility for assembling image grids using the Pillow and NumPy libraries. The code in scripts/main.py is well-structured, lacks dangerous functions (like eval or subprocess), and performs standard image processing tasks. The SKILL.md documentation includes proactive security measures, such as explicit instructions for the agent to check for path traversal (../) and strictly enforce input validation to prevent scope creep or prompt injection.