analyze-context

Analyze Context

Upload your existing files (docs, PDFs, code) and let SubAgents extract structured insights. Results are saved as clean documents - not dumped into chat context.

---

Purpose

Help users upload context that informs:

• Goal refinement
• Workspace structure suggestions
• BUILD ideas
• Integration opportunities

Key Principle:

• 01-memory/input/ is TEMPORARY - holds uploads until organized
• SubAgents write to 01-memory/input/_analysis/ (also temporary)
• The "Organize Initial Context" BUILD distributes everything to 04-workspace/
• After organization, 01-memory/input/ gets cleaned up

---

Pre-Execution

Create folders:

mkdir -p 01-memory/input/           # Where user uploads files (temporary)
mkdir -p 01-memory/input/_analysis/ # Where analysis results go (temporary)

---

Workflow

Step 1: Explain & Invite Upload

Display:

CONTEXT UPLOAD
----------------------------------------------------

Have files that show what you work on?

I'll analyze them and extract:
  - What you do (role, domain)
  - Patterns in your work
  - Tools you use (integration opportunities)
  - Ideas for what to build

Upload to:
→ 01-memory/input/

Supported:
  - PDFs (text extraction automatic)
  - Word docs (.docx)
  - Text files (.txt, .md)
  - Images (.png, .jpg - for screenshots)
  - Code files

When done, type 'done'.

Wait for user to upload and confirm "done"

---

Step 2: Scan & Validate Files

Use Glob tool to scan 01-memory/input/*:

from pathlib import Path

input_dir = Path("01-memory/input/")
uploaded_files = [f for f in input_dir.glob("*") if f.is_file()]

if not uploaded_files:
    print("No files found in 01-memory/input/")
    print("Please upload files and type 'done' when ready.")
    return

# Calculate total size
total_kb = sum(f.stat().st_size for f in uploaded_files) / 1024
file_count = len(uploaded_files)

print(f"Found {file_count} files ({total_kb:.1f} KB)")

---

Step 3: Pre-Processing (Extract & Normalize)