claude-memFind the decision.
Use Smart search (default) with mem-search: search the index, inspect nearby timeline context, then read only the observations that matter.
Memory skills
Find it. Understand it. Build on it.
Use the bundled claude-mem skills to search your history, plan work, explore a codebase, and build focused knowledge from past sessions.
In practice
Useful moves for the next session.
Explore the code.
Use smart-explore to inspect symbols and structure before opening full files. The amount of context saved depends on the task.
Make a plan.
Use make-plan to break the task into work the agent can carry out, with past decisions available as context.
Bundled toolkit
There’s a skill for that.
mem-searchSmart search: index, timeline, then selected details
knowledge-agentBuild a queryable brain from corpora
smart-exploreInspect code structure before reading full files
learn-codebaseDeep-learn a repo into memory
make-planResearch and write implementation plans
doOrchestrate plan execution with subagents
pathfinderTrace code paths and flows
timeline-reportReport what happened across sessions
babysitSupervise a long-running agent task
how-it-worksExplain how claude-mem works
version-bumpBump versions and changelog
Commands and availability depend on your installed plugin version and client. Check the current documentation for installation and usage.
Keep the useful part
Your next session
starts with context.
Install the open-source engine. Add CMEM Pro when you want hosted observation and cloud sync.