CausaMem 0.17（因果记忆 / 认知锚定）
↓
self-improving（执行质量 / 复盘固化）
↓
proactivity（前瞻行为 / 主动跟进）
↓
Beads（执行状态 / 审计轨迹，作为 R0 现实源回流 CausaMem）
| System | Key Question | When |
|---|---|---|
| CausaMem | "What facts, rules, decisions, and causal chains should anchor this judgment?" | Before judgment / task start |
| self-improving | "What did this teach me, and what should be fixed next time?" | Task end / after failure |
| proactivity | "What should I do next proactively?" | Idle / periodic / follow-through |
| Beads | "What is the current execution state and audit trail?" | During execution |
- CausaMem is the primary memory and judgment system. It owns long-term causal memory, the R0/C1/I2 cognitive backbone, the expanded 8-layer anchoring stack, and deterministic write gates.
- self-improving does not replace memory. It stores execution lessons, corrections, domain heuristics, and quality improvements.
- proactivity does not replace memory. It stores active follow-through state, proactive boundaries, heartbeat rules, and next-action momentum.
- Beads does not replace CausaMem. It is an execution-tracking reality source for tasks, dependencies, ready queues, and audit traces.
CausaMem -> recalls C1/R0 context before judgment
self-improving -> solidifies lessons after execution
proactivity -> drives the next useful action
Beads -> feeds execution facts back as R0 evidence
Short form:
CausaMem = brain
self-improving = learning muscle
proactivity = action nerve
Beads = execution sensor
Cognition structures raw facts into judgment-ready memory. Intuition injects compressed judgment anchors before reasoning. Causality links evidence, decisions, outcomes, conflicts, and validity into traceable cause-effect chains.
R0 Reality Evidence
F1 Atomic Factlets
S2 Refined Summaries
P3 Profiles and Scenes
K4 Wiki Knowledge
D5 Dream Consolidation
C6 Causal Chains
I7 Intuition Injection