| name | automem |
| description | Persistent AutoMem memory via the native AutoMem OpenClaw plugin tools. |
| user-invocable | true |
| metadata | {"openclaw":{"skillKey":"automem"}} |
AutoMem
Use the native AutoMem tools exposed by the AutoMem plugin.
Natural language mappings
remember ... or store this -> call automem_store_memory
what do you know about ... or recall ... -> call automem_recall_memory
update memory ... -> call automem_update_memory
delete memory ... -> recall first when needed, then call automem_delete_memory
link these memories ... -> call automem_associate_memories
is memory healthy? -> call automem_check_health
Slash command behavior
Treat /automem remember ..., /automem recall ..., /automem update ..., and /automem delete ... as direct requests to use the matching AutoMem tool flow above.
Rules
- Recall first for prior decisions, preferences, ongoing projects, and debugging history.
- Store durable outcomes: decisions, bug fixes, patterns, preferences, and important context.
- Keep content compact:
Brief title. Context and details. Impact/outcome.
- If deletion is ambiguous, recall candidates first and ask for confirmation with ids before deleting.
- Use
memory-core alongside AutoMem when file-backed workspace memory is helpful. It complements AutoMem; it is not a replacement.