
Memory
- 349 installs
- 24 repo stars
- Updated December 19, 2025
- johnlindquist/claude
memory is an agent skill that stores and retrieves project knowledge across sessions using the basic-memory CLI and semantic search for developers who need coding agents to recall facts, preferences, and decisions.
About
memory is an agent skill from johnlindquist/claude that provides persistent knowledge storage across coding sessions using the basic-memory CLI. After pip install basic-memory, agents write notes with basic-memory tool write-note—including title, markdown content, and optional tags—and retrieve them via semantic search to rebuild topic context between sessions. Developers reach for memory when project facts, architectural decisions, or preferences must survive session boundaries without re-prompting from scratch each time. The skill supports folder-targeted notes, tagged retrieval, and topic-based context building for long-running repositories where volatile context windows lose institutional knowledge. memory complements file-based planning skills by storing unstructured recall policies and searchable knowledge rather than phased task plans.
- Session vs long-term memory
- Retrieval and ranking
- Write/update policies
- Privacy and scope boundaries
- Conflict resolution
Memory by the numbers
- 349 all-time installs (skills.sh)
- +5 installs in the week ending Aug 2, 2026 (Skillselion tracking)
- Ranked #2,139 of 16,546 AI & Agent Building skills by installs in the Skillselion catalog
- Data as of Aug 2, 2026 (Skillselion catalog sync)
npx skills add https://github.com/johnlindquist/claude --skill memoryAdd your badge
Show developers this skill is listed on Skillselion. Paste this into your README.
| Installs | 349 |
|---|---|
| repo stars | ★ 24 |
| Last updated | December 19, 2025 |
| Repository | johnlindquist/claude ↗ |
How do you give coding agents persistent memory across sessions?
Design persistent memory stores and recall policies so coding agents retain project facts, preferences, and decisions across sessions.
Who is it for?
Developers who want coding agents to retain project facts, preferences, and decisions across sessions using basic-memory semantic search instead of re-explaining context.
Skip if: Developers who only need phased task tracking with status gates—use planning-with-files for structured multi-step plans instead of unstructured recall.
When should I use this skill?
Project facts, preferences, or decisions must be saved or semantically searched across multiple coding agent sessions.
What you get
Tagged markdown notes in basic-memory stores and semantically retrieved context bundles for future agent sessions.
- tagged knowledge notes
- semantic search results
Files
Memory - Persistent Knowledge Storage
Store and retrieve knowledge across sessions using semantic search.
Prerequisites
Install basic-memory:
pip install basic-memoryCLI Reference
Write a Note
# Basic note
basic-memory tool write-note --title "Note Title" --content "Note content in markdown"
# With tags
basic-memory tool write-note --title "React Patterns" --content "# Content here" --tags "react,patterns,frontend"
# To specific folder
basic-memory tool write-note --title "Meeting Notes" --content "# Notes" --folder "meetings"
# With specific project
basic-memory tool write-note --title "Project Notes" --content "# Notes" --project myprojectRead a Note
# By identifier/permalink
basic-memory tool read-note "note-title"
# From specific project
basic-memory tool read-note "note-title" --project myprojectSearch Memories
# Semantic search (query is positional)
basic-memory tool search-notes "your search query"
# Limit results
basic-memory tool search-notes "react hooks" --page-size 10
# Search by permalink
basic-memory tool search-notes "pattern" --permalink
# Search by title
basic-memory tool search-notes "meeting" --title
# Filter by date
basic-memory tool search-notes "feature" --after_date "7d"
# With specific project
basic-memory tool search-notes "authentication" --project myprojectBuild Context for a Topic
# Get related notes for a URL/topic (URL is positional)
basic-memory tool build-context "memory://topic/authentication"
# With depth and timeframe
basic-memory tool build-context "memory://note/my-note" --depth 2 --timeframe 30d
# With specific project
basic-memory tool build-context "memory://topic/api" --project myprojectList Recent Activity
# Default (depth 1, 7 days)
basic-memory tool recent-activity
# With depth
basic-memory tool recent-activity --depth 3
# With timeframe
basic-memory tool recent-activity --timeframe 30d
# Combined
basic-memory tool recent-activity --depth 3 --timeframe 14dContinue Conversation
# Get prompt to continue previous work
basic-memory tool continue-conversation "previous topic or context"Sync Database
basic-memory syncCheck Status
basic-memory statusNote Format
Notes are stored as markdown with YAML frontmatter:
---
title: My Note
tags: [tag1, tag2]
created: 2024-01-15
---
# My Note
Content here in markdown format.
## Sections
More content...Workflow Patterns
Save Learning
When you discover something useful:
basic-memory tool write-note \
--title "TypeScript Utility Types" \
--content "# Utility Types\n\n- Partial<T> - Makes all properties optional\n- Required<T> - Makes all properties required\n- Pick<T, K> - Picks specific properties" \
--tags "typescript,types,reference"Save Decision
When making an architectural decision:
basic-memory tool write-note \
--title "Auth Strategy Decision" \
--content "# Decision: Use JWT for API auth\n\n## Context\n...\n\n## Decision\n...\n\n## Consequences\n..." \
--tags "architecture,auth,decision" \
--folder "decisions"Retrieve Context
Before starting related work:
# Search for relevant notes
basic-memory tool search-notes "authentication jwt tokens"
# Or build comprehensive context
basic-memory tool build-context "memory://topic/api-authentication"Review Recent Work
basic-memory tool recent-activity --depth 5 --timeframe 7dUse Cases
Personal Wiki
- Store code snippets and patterns
- Document project decisions
- Keep reference material
Learning Log
- Record things you learn
- Tag by topic for later retrieval
- Build knowledge over time
Project Context
- Save project-specific knowledge
- Retrieve relevant context at session start
- Share knowledge across sessions
Best Practices
1. Use meaningful titles - They become permalinks 2. Add tags - Improves search and organization 3. Use markdown - Full markdown support in content 4. Organize with folders - Group related notes 5. Search before writing - Check if knowledge already exists 6. Keep notes focused - One topic per note 7. Update, don't duplicate - Revise existing notes
Integration Pattern
Session Start
# Get context for today's work
basic-memory tool build-context "memory://topic/feature-youre-building"During Work
# Save discoveries
basic-memory tool write-note --title "Discovery Title" --content "What I learned"Session End
# Sync to ensure persistence
basic-memory syncRelated skills
FAQ
What CLI does the memory skill use?
memory uses the basic-memory CLI installed via pip install basic-memory. Agents write notes with basic-memory tool write-note (title, markdown content, optional tags) and retrieve knowledge through semantic search across sessions.
How is memory different from planning-with-files?
memory stores unstructured project knowledge and recall policies via basic-memory semantic search. planning-with-files maintains structured phased plans in task_plan.md, findings.md, and progress.md with completion gates.