clawdhub install agent-memory

from src.memory import AgentMemory

mem = AgentMemory()

# Remember facts
mem.remember("Important information", tags=["category"])

# Learn from experience
mem.learn(
    action="What was done",
    context="situation",
    outcome="positive",  # or "negative"
    insight="What was learned"
)

# Recall memories
facts = mem.recall("search query")
lessons = mem.get_lessons(context="topic")

# Track entities
mem.track_entity("Name", "person", {"role": "engineer"})

## Memory Protocol

On session start:
1. Load recent lessons: `mem.get_lessons(limit=5)`
2. Check entity context for current task
3. Recall relevant facts

On session end:
1. Extract durable facts from conversation
2. Record any lessons learned
3. Update entity information