Elite Longterm Memory
Give coding agents durable memory across sessions with WAL, vectors, and git-backed knowledge so context survives restarts.
Overview
Elite Longterm Memory is a journey-wide agent skill that layers WAL, vector search, and git-based knowledge graphs so solo builders never lose agent context across sessions.
Install
npx skills add https://github.com/nextfrontierbuilds/elite-longterm-memory --skill elite-longterm-memoryWhat is this skill?
- Combines 6 approaches: WAL protocol, LanceDB vectors, git-notes knowledge graph, cloud backup, and hygiene
- Tiered architecture: hot SESSION-STATE.md, warm vector store, cold git-based graph
- Targets Clawdbot, Moltbot, Claude, and GPT-style agents that lose context between runs
- Requires memory-lancedb plugin and OPENAI_API_KEY per skill metadata
- Version 1.2.0—bulletproof write-ahead logging to avoid losing decisions
- Combines 6 proven memory approaches in one architecture
- Skill version 1.2.0
Adoption & trust: 2.3k installs on skills.sh; 13 GitHub stars; 3/3 security scanners passed (skills.sh audits).
What problem does it solve?
Your agent forgets decisions after each session, so you rehash architecture, prefs, and mistakes instead of shipping.
Who is it for?
Solo builders running persistent coding agents with LanceDB/memory plugins and OpenAI embeddings configured.
Skip if: One-off tasks, teams forbidding git-notes or external API keys, or users who only need a single markdown PROJECT.md.
When should I use this skill?
Setting up or maintaining bulletproof long-term memory for AI agents that must retain context across sessions.
What do I get? / Deliverables
You run a tiered memory system with durable WAL, semantic recall, and git-backed notes so agents resume with full project context.
- Tiered hot/warm/cold memory layout (SESSION-STATE, vectors, git-notes)
- WAL-backed persistence and maintenance hygiene routines
Recommended Skills
Journey fit
Useful at every journey phase - explore requirements and options before committing to a direction.
Where it fits
Recall prior scope decisions and user quotes when revisiting PRD sections days later.
Index implementation choices into LanceDB so the agent picks consistent patterns across backend tasks.
Pull past review findings from git-notes graph before merging a follow-up PR.
Store incident timelines in durable memory so on-call agent sessions continue from the last WAL checkpoint.
How it compares
Agent memory orchestration skill—not a hosted SaaS memory product or a simple SESSION.md snippet.
Common Questions / FAQ
Who is elite-longterm-memory for?
Indie builders and agent operators using Claude, GPT, Clawdbot, or Moltbot who need cross-session recall, vector search, and durable decision logs.
When should I use elite-longterm-memory?
Whenever long-running agent work spans phases—scoping in Validate, implementing in Build, debugging in Operate— and you need the same context without re-prompting from scratch.
Is elite-longterm-memory safe to install?
It requests API keys and filesystem or plugin access; review Security Audits on this page and limit secrets scope before enabling cloud backup or vector indexing.
SKILL.md
READMESKILL.md - Elite Longterm Memory
# Elite Longterm Memory 🧠 **The ultimate memory system for AI agents.** Combines 6 proven approaches into one bulletproof architecture. Never lose context. Never forget decisions. Never repeat mistakes. ## Architecture Overview ``` ┌─────────────────────────────────────────────────────────────────┐ │ ELITE LONGTERM MEMORY │ ├─────────────────────────────────────────────────────────────────┤ │ │ │ ┌─────────────┐ ┌─────────────┐ ┌─────────────┐ │ │ │ HOT RAM │ │ WARM STORE │ │ COLD STORE │ │ │ │ │ │ │ │ │ │ │ │ SESSION- │ │ LanceDB │ │ Git-Notes │ │ │ │ STATE.md │ │ Vectors │ │ Knowledge │ │ │ │ │ │ │ │ Graph │ │ │ │ (survives │ │ (semantic │ │ (permanent │ │ │ │ compaction)│ │ search) │ │ decisions) │ │ │ └─────────────┘ └─────────────┘ └─────────────┘ │ │ │ │ │ │ │ └────────────────┼────────────────┘ │ │ ▼ │ │ ┌─────────────┐ │ │ │ MEMORY.md │ ← Curated long-term │ │ │ + daily/ │ (human-readable) │ │ └─────────────┘ │ │ │ │ │ ▼ │ │ ┌─────────────┐ │ │ │ SuperMemory │ ← Cloud backup (optional) │ │ │ API │ │ │ └─────────────┘ │ │ │ └─────────────────────────────────────────────────────────────────┘ ``` ## The 5 Memory Layers ### Layer 1: HOT RAM (SESSION-STATE.md) **From: bulletproof-memory** Active working memory that survives compaction. Write-Ahead Log protocol. ```markdown # SESSION-STATE.md — Active Working Memory ## Current Task [What we're working on RIGHT NOW] ## Key Context - User preference: ... - Decision made: ... - Blocker: ... ## Pending Actions - [ ] ... ``` **Rule:** Write BEFORE responding. Triggered by user input, not agent memory. ### Layer 2: WARM STORE (LanceDB Vectors) **From: lancedb-memory** Semantic search across all memories. Auto-recall injects relevant context. ```bash # Auto-recall (happens automatically) memory_recall query="project status" limit=5 # Manual store memory_store text="User prefers dark mode" category="preference" importance=0.9 ``` ### Layer 3: COLD STORE (Git-Notes Knowledge Graph) **From: git-notes-memory** Structured decisions, learnings, and context. Branch-aware. ```bash # Store a decision (SILENT - never announce) python3 memory.py -p $DIR remember '{"type":"decision","content":"Use React for frontend"}' -t tech -i h # Retrieve context python3 memory.py -p $DIR get "frontend" ``` ### Layer 4: CURATED ARCHIVE (MEMORY.md + daily/) **From: Clawdbot native** Human-readable long-term memory. Daily lo