From rok-memory
Encrypted decentralized memory layer backed by Fileverse via the rok-mcp server. Proactively loads memories at conversation start if autoLoad is configured. Triggers on: "list/read/write/grant memories", "store this at /scope", "grant access", "remember this", "rok setup", or any memory CRUD operation.
How this skill is triggered — by the user, by Claude, or both
Slash command
/rok-memory:rok-memoryThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
All operations go through `rok_memory:*` MCP tools.
All operations go through rok_memory:* MCP tools.
| Tool | Role | Purpose |
|---|---|---|
rok_memory:setup | any | Configure credentials and write ~/.rok/session.json |
rok_memory:login | any | Start session (if not using setup/auto-config) |
rok_memory:logout | any | End session, zeroize credentials |
rok_memory:list | any | List all accessible memories |
rok_memory:read | any | Decrypt a memory by scope + key |
rok_memory:write | owner | Encrypt and store a memory |
rok_memory:grant | owner | Export a scoped read key for delegation |
rok_memory:propose | any | Stage (agent) or accept (owner) a proposed write |
rok_memory:sync | any | Batch upsert with dedup — owner writes, agent proposes |
rok_memory:status | any | Show session role, scope, memory count |
When the user says "remember this", "store this", "save to memory", or similar:
rok_memory:write with scope and keyrok_memory:propose to stage the changeWhen autoLoad is enabled in ~/.rok/session.json, a Stop hook fires at conversation end.
It injects current memories and prompts you to call rok_memory:sync with any new knowledge.
/finance can read /finance/q1 but not /engineeringrok://memory/{scope}/{key}npx claudepluginhub manojkgorle/ethmumbai-rok --plugin rok-memoryCross-host durable memory for AI agents using the ling-mem CLI. Maintains a three-tier model of who the user is across sessions and hosts (Claude Code, Codex, OpenClaw).
Invoked via /memsy slash command; classifies intent (search, store, switch profile, list, doctor, setup) and runs the matching Memsy workflow for context memory.
Persistent memory for AI agents via hosted MCP. Remember decisions, recall project context, manage TODOs, and govern memory across sessions and tools.