| name | memcontext |
| description | Use MemContext memory tools to recall prior decisions and store durable user or project facts. |
| user-invocable | false |
| metadata | {"openclaw":{"emoji":"🧠","requires":{"config":["plugins.entries.openclaw-memcontext.enabled"]}}} |
MemContext
Use this skill when long-term context matters across OpenClaw sessions.
What MemContext is for
- Recall prior decisions, preferences, facts, and project context.
- Persist durable information that should survive resets, compaction, and future sessions.
- Keep memory focused on stable information, not transient chatter.
When to search memory
- At the start of a conversation when the user references prior context.
- Before answering questions like "what did we decide", "remember this", "how do I prefer", or "what stack are we using".
- Before making architecture or workflow decisions that might already have been decided.
Use the memcontext_search tool for manual lookups when the current prompt context is not enough.
When to save memory
Use memcontext_store immediately when the user shares durable information such as:
- stable preferences or dislikes
- important personal or team facts
- project conventions or architecture decisions
- durable context that will matter in later sessions
Do not save one-off debugging noise, temporary outputs, or generic chit-chat.
Avoid saving secrets, tokens, passwords, credential URLs, or other sensitive data.
Categories
preference for likes, dislikes, and workflow preferences
fact for durable truths about the user or project
decision for chosen tools, architecture, or conventions
context for important recent background that should survive session boundaries
Examples
- Save: "User prefers pnpm over npm" ->
preference
- Save: "This project uses Hono for the API" ->
fact
- Save: "We chose MemContext as the OpenClaw memory backend" ->
decision
- Save: "We are currently building the OpenClaw plugin in a separate repository" ->
context