output-workspace-cleanup

Output Workspace Cleanup

When to use

Load this skill when:

• The output directory (~/.Hermes/workspace/output/ or similar) has 10+ project folders at the root level with no categorical grouping
• Project folder names are inconsistent (some use video ID, some use topic, some use auto-generated yt-dlp names)
• You need to separate final deliverables from intermediate working files
• The user says "整理一下" (tidy up) or "结构更清晰" regarding saved outputs
• A new project type emerges and needs a home in the existing structure

Core workflow

Phase 1: Survey

import os

output = os.path.expanduser("~/.Hermes/workspace/output/")
folders = sorted([d for d in os.listdir(output) if os.path.isdir(os.path.join(output, d)) and not d.startswith('.')])

# Categorize each folder by:
# 1. Contents — does it have PDF reports, MP4 videos, scripts, transcripts, etc.?
# 2. Nature — is it a report, a video production, a download, infographic, or system file?

Key patterns to identify:

Has	Likely category
PDF consulting-style reports + markdown + transcripts	reports/consulting/
Bilingual subtitle files + multiple MP4 versions	video_projects/bilingual_subtitle/
Single MP4 download + maybe SRT	video_projects/raw/ or x_videos/ (if X/Twitter)
Python scripts for KOL/influencer research	reports/influencer_research/
Infographic generator scripts + images	infographics/misc/
Short video (< 3 min)	video_projects/shorts/
Update logs, cron files	system/maintenance/

Phase 2: Design structure

Standard top-level schema:

output/
├── reports/
│   ├── consulting/            # Consulting-style PDF reports from video analysis
│   └── influencer_research/   # KOL/influencer investigation reports
├── video_projects/
│   ├── bilingual_subtitle/    # Bilingual subtitle video productions
│   ├── shorts/                # Short video downloads
│   └── raw/                   # Raw source videos (IELTS, training, etc.)
├── x_videos/                  # Shared X/Twitter video pool (keep flat)
├── infographics/
│   └── misc/
├── system/
│   └── maintenance/           # Agent update logs, cron job records
└── cache/
    └── skill_repo/            # Cached skill repositories