# Knowledge Base — PARA System

Structured knowledge management based on Thiago Forte's PARA method. This replaces ad-hoc memory files as the primary knowledge store for all agents.

## Structure

### projects/
Active projects with clear deadlines or deliverables. One markdown file per project.
- `dfw-launch.md` — DFW market entry project tracker ✅ exists

Future (create when needed):
- `leadspanther-site.md` — Website and DeckEngine
- `mission-control.md` — Dashboard

### areas/
Ongoing areas of responsibility with no end date. Updated continuously.
- `agent-operations.md` — Agent specialization, skill incubator, session management ✅ exists

Future (create when needed):
- `sales-pipeline.md` — Active prospects, deal stages, follow-ups
- `infrastructure.md` — Server, cron, credentials, system health
- `finances.md` — Revenue, costs, runway, funding status

### resources/
Reference material. Rarely changes. Indexed for fast lookup.
- `operating-principles.md` — 25 operating principles from transcript analysis ✅ exists

Future (create when needed):
- `frameworks.md` — Sales frameworks (SPIN, Challenger, MEDDIC, etc.)
- `target-verticals.md` — DFW vertical research (Rene's output post-funding)
- `tools-and-apis.md` — Tool configs, API docs, integration notes

### archive/
Completed projects and outdated resources. Moved here, never deleted.

## Usage Rules

1. **Every agent** should check `knowledge/` before starting work on a topic
2. **Daily notes** (`memory/YYYY-MM-DD.md`) remain raw session logs
3. **Knowledge base** is the curated, structured version — distilled from daily notes
4. **Nightly consolidation** (2 AM cron) extracts important info from daily notes → updates knowledge base
5. **MEMORY.md** remains the top-level summary for main sessions — knowledge/ is the detailed backing store
6. **When you learn something important**, update the relevant knowledge file immediately — don't wait for consolidation