# Resources

## Primary Sources

1. **Samuel's META2 Antworten** (local, highest trust)
   - `frage_05_qualitaet/ANTWORT.md` — Quality metrics for agent work
   - `frage_07_ueberforderung/ANTWORT.md` — Overload and parallel system breaks
   - `frage_01_kontrolle/ANTWORT.md` — Control and meta-management
   - `frage_04_thalamus/ANTWORT.md` — Digital thalamus / filtering

2. **Hermes Agent Documentation** (local)
   - `~/TH-Mannheim/` and `~/Projekte/Agent-Baukasten/` — existing agent architectures
   - Hermes skills and registry system

## External Sources (to evaluate)

- [ ] Andy Grove, "High Output Management" — output-focused management, OKRs
- [ ] Kanban/WIP limits literature — vertical depth vs horizontal breadth
- [ ] Software craftsmanship: "Definition of Done" patterns
- [ ] Agent evals: finding reliable quality metrics for LLM pipelines

## Trust Levels

- **High**: Samuel's own META2 analysis (self-generated, self-verified)
- **Medium**: Hermes architecture docs (designed but not all operationalized)
- **Low**: General productivity/management literature (needs filtering through Samuel's actual context)