Setup & Installation
Or with the ClawHub CLI, for registry-managed skill folders outside a full OpenClaw workspace:
Version History
- Updated the description to clarify the "remember-this" trigger and expanded on memory use cases (preferences, goals, boundaries, prior work, decisions, dates, and todos). - No substantial changes to functionality or rules in the skill documentation.
What This Skill Does
Tracks user preferences, goals, decisions, and project context across conversations. Entries are classified as facts, preferences, goals, or hypotheses, stored in daily notes and a long-term profile file. Hypotheses decay if unvalidated; durable items are promoted based on repetition and impact.
Without structured memory, users repeat context every session; this skill maintains a typed, decay-aware profile that persists and evolves with explicit validation rules.
When to use it
- Recalling a preferred coding style in future sessions
- Storing a stated boundary like 'never suggest rewrites'
- Tracking ongoing project priorities across multiple conversations
- Surfacing earlier decisions when context becomes relevant again
- Personalizing response tone based on known user behavior
Example Workflow
Here's how your AI assistant might use this skill in practice.
User says: "Remember I prefer short answers and no bullet lists"
- 1Classifies the signal as a PREFERENCE
- 2Appends a timestamped entry to the current daily memory file with impact score 2
- 3Promotes to long-term MEMORY.md immediately because it is an explicit preference
- 4Adapts all future responses to match the stored preference
Preference stored in MEMORY.md and applied to subsequent interactions