Scan Report
0 /100
arch-video-cut
Automatic Architecture Video Editing Workflow with Self-Learning Preferences
Legitimate architecture video editing workflow using ffmpeg for merging, subtitling, and audio mixing with a self-learning preference system. All shell execution is documented and necessary for video processing.
Safe to install
No action needed. Skill is safe to use.
| Resource | Declared | Inferred | Status | Evidence |
|---|---|---|---|---|
| Shell | WRITE | WRITE | ✓ Aligned | scripts/full_workflow.py:46 subprocess.run(..., shell=True) — required for ffmpe… |
| Filesystem | WRITE | WRITE | ✓ Aligned | All writes scoped to data/ and config/ directories as documented |
1 findings
Medium External URL 外部 URL
http://www.ual-studio.com/ SKILL.md:242 File Tree
6 files · 29.5 KB · 1018 lines Markdown 3f · 510L
Python 3f · 508L
├─
▾
scripts
│ ├─
full_workflow.py
Python
│ ├─
manage_preferences.py
Python
│ └─
preference_learner.py
Python
├─
README.md
Markdown
├─
SELF_LEARNING_GUIDE.md
Markdown
└─
SKILL.md
Markdown
Security Positives
✓ All subprocess shell execution is directly documented via ffmpeg installation instructions in SKILL.md
✓ File I/O is strictly scoped to data/ and config/ directories, no sensitive paths accessed
✓ No external network connections made by runtime code (external URL is only in author branding in SKILL.md)
✓ No credential harvesting, environment variable iteration, or sensitive file access
✓ No obfuscation, base64 execution, or anti-analysis techniques
✓ No hidden functionality — SKILL.md accurately describes all capabilities
✓ No supply chain risks — uses only standard library and ffmpeg (documented as a prerequisite)
✓ Self-learning system only writes user preferences to config/user_preferences.json, no exfiltration