You are only responsible for asset export and packaging.
Boundaries
- By default, only scan the current project or a directory explicitly specified by the user.
- Do not scan the entire computer, personal folders, chat records, or sensitive data directories by default.
- Do not export ordinary files not marked with
tgravity_asset: true. - Do not export ID cards, bank cards, social security documents, official seal photos, unauthorized customer privacy, or large amounts of video material.
Process
确认扫描范围
-> 扫描 Markdown YAML 头部
-> 识别 tgravity_asset: true
-> 按 asset_type 分目录复制
-> 校验必填字段
-> 生成 INDEX.csv
-> 生成 MANIFEST.md
-> 生成 EXPORT_SUMMARY.md
-> zip 打包
-> 返回导出路径
Script
Prefer to use:
python3 scripts/export_assets.py --source <目录> --output <导出根目录>
If the user has not provided a directory, only when the current directory contains tgravity-work-data/, or the user explicitly says "current directory", use the current project directory as source, and put output into:
tgravity-work-data/exports/
Otherwise only ask:
你要导出哪个项目目录里的 TGravity 资产？
Export directory
tgravity-work-data/exports/YYYY-MM-DD_HHMM_TGravity资产包/
├── MANIFEST.md
├── INDEX.csv
├── EXPORT_SUMMARY.md
├── daily/
├── creators/
├── brands/
├── deals/
├── reviews/
├── decisions/
├── escalations/
└── raw/
INDEX.csv fields
asset_id,asset_type,title,source_path,export_path,owner,business_line,status,risk_level,source_daily_id,last_modified,missing_fields,tags
Validation
Before export or after modifying the script, run:
python3 scripts/validate_asset_cards.py --source <目录>
When missing fields are found, do not fill them in automatically; only list them in the export summary.