Multi-Agent Ralph Loop
github.com/alfredolopez80/multi-agent-ralph-loop
What it is
Section titled “What it is”Autonomous orchestration framework for Claude Code with MemPalace-inspired memory (4-layer L0-L3 stack), 818-token “wake-up” context, parallel-first Agent Teams (6 teammates), Aristotle First Principles methodology, 4-stage quality gates, 22 lifecycle hooks, 925+ tests.
The 4 layers (per their docs)
Section titled “The 4 layers (per their docs)”- L0: hot context this turn
- L1: session state
- L2: mission knowledge
- L3: cross-mission Obsidian-style knowledge graph
What we kept
Section titled “What we kept”- 4-layer memory model. Borrowed verbatim. Our equivalent: L0 runtime context, L1 raw.md, L2 summary.md/MEMORY.md, L4 (we kept the cross-mission layer at L4 to leave room)
identity/<role>.md. - Lifecycle hooks pattern. They have 22; we have 12. Different events but same structural idea.
What we dropped
Section titled “What we dropped”- Obsidian vault knowledge graph. Markdown files in directories work fine. The graph format adds complexity without payoff at our scale.
- Aristotle First Principles methodology. We didn’t adopt the explicit framing. Decisions are tracked in
decisions/log.mddirectly. - 818-token “wake-up” hard cap. We have a 800-token cap on
summary.md(close enough) but no special “wake-up” file.
Differences in philosophy
Section titled “Differences in philosophy”| Ralph Loop | Papercup |
|---|---|
| Obsidian-style cross-mission graph | Per-role flat markdown files |
| Aristotle First Principles framing | Decision log without framing layer |
| 22 lifecycle hooks | 12 lifecycle hooks |
| 6 fixed teammates | 13 customisable role prompts |
Honest take
Section titled “Honest take”The 4-layer memory model is one of the genuinely original ideas in the ecosystem. We adopted it. The Obsidian/graph layer is more ceremony than we wanted but the pattern is sound — if you have hundreds of missions and need cross-link discoverability, the graph format pays off.