file-processing

name	File Processing
description	Process, classify, organize, and index documents. Extracts relevant information to user's memory.
allowed-tools	Bash, Read, Write, Glob

File Processing Skill

When asked to "process", "organize", or "file" documents, follow this complete workflow.

---

The Workflow

1. Read & Understand

Read each document to understand its content:

• PDFs: Use Read tool (supports PDF extraction)
• Images: Use Read tool (multimodal)
• Text files: Use Read tool

Extract key information:

• Document type (ID, legal, medical, business, personal)
• People involved
• Dates and events
• Important facts

2. Classify & Organize

Move files to the appropriate folder based on content:

Document Type	Destination
Personal ID (cédula, passport, police certs)	/home/jpelaez/documents/personal/id/
Military records	/home/jpelaez/documents/personal/military/
Family member docs	/home/jpelaez/documents/personal/{name}/
Legal (divorces, contracts)	/home/jpelaez/documents/legal/{category}/
Migration/immigration	/home/jpelaez/documents/legal/migration/
Medical records, billing	/home/jpelaez/documents/medical/
Company documents	/home/jpelaez/{CompanyName}/documents/
Contact/third-party docs	/home/jpelaez/documents/contacts/{name}/

Important: Company documents go in the company folder, not personal documents.

3. Index Files

Use file_indexer.py to add files to the searchable index:

$LOLABOT_HOME/tools/files.sh scan /path/to/folder --tags "tag1,tag2"
$LOLABOT_HOME/tools/files.sh add "/path/to/file.pdf" -d "Description" -t "tags"

Or with full command:

source $LOLABOT_HOME/.venv/bin/activate
python $LOLABOT_HOME/tools/file_indexer.py scan /path --tags "tags"

Tagging guidelines:

• Always include document type: personal, legal, medical, business
• Include person name if relevant: juan, felipe, marco
• Include category: id, passport, divorce, migration

4. Extract to Memory

For documents containing personal information about Juan or his family, add relevant facts to memory:

$LOLABOT_HOME/tools/memory.sh add "Fact extracted from document" --type fact --tags "relevant,tags"

Or with full command:

source $LOLABOT_HOME/.venv/bin/activate
python $LOLABOT_HOME/tools/memory_indexer.py add "..." --type fact --tags "..."

What to extract:

Document Contains	Memory Type	Example
Birth date, ID numbers	fact	"Juan's cédula is 79,672,150"
Events with dates	event	"Felipe born December 26, 2007"
Relationships	person	"Jose Miguel Nunes Pelaez is Juan's cousin"
Addresses	fact	"Current address: 2852 Kalmia Ave..."
Financial info	fact	"BCH medical debt: $12,171.78"
Expiration dates	fact	"Felipe's passport expires Feb 14, 2026"

Don't extract:

• Redundant information already in memory
• Trivial details
• Sensitive credentials (store securely elsewhere)

---

Folder Structure Reference

/home/jpelaez/
├── documents/
│   ├── personal/
│   │   ├── id/           # Juan's ID documents
│   │   ├── military/     # Military records
│   │   ├── felipe/       # Son's documents
│   │   ├── marco/        # Father's documents
│   │   └── {family}/     # Other family members
│   ├── legal/
│   │   ├── divorces/
│   │   ├── migration/
│   │   └── contracts/
│   ├── medical/
│   └── contacts/
│       └── {name}/       # Third-party documents
├── 3Metas/
│   └── documents/        # 3Metas company docs
├── PPM/
│   └── documents/        # PPM company docs
└── {Company}/
    └── documents/        # Other company docs

---

Transport Folder

New documents typically arrive in /srv/fileserver/transport/ (local fileserver on mini-lola).

When processing transport:

1. List all files in transport
2. Process each file through the workflow
3. After successful copy and indexing, originals can remain or be removed

---

Example Session

# 1. List what's in transport
ls /mnt/fileserver/transport/

# 2. Read a PDF
# (Use Read tool on each file)

# 3. Create destination folder if needed
mkdir -p /home/jpelaez/documents/personal/felipe

# 4. Copy file to destination
cp "/mnt/fileserver/transport/Felipe Birth Certificate.pdf" \
   "/home/jpelaez/documents/personal/felipe/"

# 5. Index the file
source $LOLABOT_HOME/.venv/bin/activate
python $LOLABOT_HOME/tools/file_indexer.py scan \
  /home/jpelaez/documents/personal/felipe --tags "personal,felipe,family"

# 6. Add to memory
python $LOLABOT_HOME/tools/memory_indexer.py add \
  "Felipe Pelaez born December 26, 2007. Mother: Claudia Patricia Huertas Diaz" \
  --type fact --tags "family,felipe"

---

Quick Reference

Action	Command
Scan folder	files.sh scan /path --tags "tags"
Add single file	files.sh add "/path" -d "desc" -t "tags"
Search files	files.sh find "query"
Add memory	memory.sh add "fact" --type TYPE --tags "tags"
Search memory	memory.sh find "query"