memory
Manage long-term memories using date-based files. Allows creating, listing, and reading memories.
When & Why to Use This Skill
The Memory skill provides a robust framework for managing long-term information through a structured, date-based file system. It empowers AI agents to persist critical context, architectural decisions, and historical data, significantly enhancing continuity and decision-making capabilities across multiple sessions and complex projects.
Use Cases
- Documenting Architectural Decisions: Store and retrieve the rationale behind technical choices to maintain consistency throughout the software development lifecycle.
- Recording Lessons Learned: Log insights from debugging or project challenges to build a searchable knowledge base that prevents the recurrence of past errors.
- Project Context Persistence: Maintain a continuous thread of project progress, milestones, and state across different chat sessions, ensuring the agent remains aligned with long-term goals.
- Tag-based Knowledge Retrieval: Organize memories with specific tags to enable rapid filtering and efficient access to relevant historical information during task execution.
| name | memory |
|---|---|
| description | Manage long-term memories using date-based files. Allows creating, listing, and reading memories. |
Memory Skill
This skill allows you to manage long-term memories in docs/memories/.
It uses the scripts/memory.py utility.
Commands
Create Memory
Create a new memory file.
Command: python3 scripts/memory.py create "<title>" "<content>" [--tags "tag1,tag2"] [--format json]
List Memories
List recent memories, optionally filtered by tag.
Command: python3 scripts/memory.py list [--tag <tag>] [--limit <limit>] [--format json]
Read Memory
Read a specific memory file.
Command: python3 scripts/memory.py read <filename_or_slug> [--format json]
Usage Instructions
- Use
createto store important architectural decisions, lessons learned, or long-term context. - Use
listto recall past memories. - Use
readto get the full content of a memory. - Memories are stored as Markdown files with YAML frontmatter.