Persistent memory patterns for cross-session learning and context retention
Implement persistent memory patterns for AI agents using ReasoningBank.
# Store a pattern
npx agentic-flow@alpha memory store "api:auth" "OAuth2 with JWT"
# Retrieve a pattern
npx agentic-flow@alpha memory get "api:auth"
# Search patterns
npx agentic-flow@alpha memory search "authentication"
# List all patterns
npx agentic-flow@alpha memory list --namespace project
| Namespace | Purpose | TTL |
|---|---|---|
session | Current session context | Until end |
project | Project-specific learnings | Permanent |
user |
| User preferences |
| Permanent |
swarm | Swarm coordination state | Swarm lifetime |
cache | Temporary cached data | 1 hour |
# Store decision with context
npx agentic-flow@alpha memory store \
"decisions:auth-method" \
'{"choice": "JWT", "reason": "stateless, scalable", "date": "2024-01-01"}'
# Store reusable code pattern
npx agentic-flow@alpha memory store \
"patterns:error-handling" \
"try-catch with custom error classes and logging"
# Store learning from successful task
npx agentic-flow@alpha memory store \
"learnings:react-hooks" \
"useCallback for event handlers, useMemo for expensive computations"
// Store memory
mcp__claude-flow__memory_usage({
action: "store",
key: "project:architecture",
value: "microservices with event-driven communication",
namespace: "project"
})
// Retrieve memory
mcp__claude-flow__memory_usage({
action: "retrieve",
key: "project:architecture",
namespace: "project"
})
// Search memories
mcp__claude-flow__memory_search({
pattern: "auth*",
namespace: "project",
limit: 10
})
ReasoningBank provides: