Agent Memory Mcp
Run a hybrid MCP memory bank so agents persist architecture, patterns, and decisions across sessions and sync with project docs.
Overview
Agent Memory MCP is an agent skill most often used in Build (also Operate iterate, Grow content) that runs a searchable MCP memory server for architecture, patterns, and decisions across a solo builder’s project.
Install
npx skills add https://github.com/sickn33/antigravity-awesome-skills --skill agent-memory-mcpWhat is this skill?
- Hybrid memory for architecture, patterns, and decisions with searchable long-term storage
- MCP tools: memory_search (query, type, tags) and memory_write (key, type, content)
- Requires Node.js v18+, git clone into .agent/skills/agent-memory, npm install and compile
- start-server script binds memory bank to a project_id and absolute workspace path
- Designed to sync automatically with project documentation over time
- Two documented MCP capabilities: memory_search and memory_write
- Node.js v18+ required for install and compile
Adoption & trust: 991 installs on skills.sh; 40.1k GitHub stars; 1/3 security scanners passed (skills.sh audits).
What problem does it solve?
Your agent forgets prior architecture choices and patterns every session, forcing you to re-paste context or hunt through scattered markdown.
Who is it for?
Indie developers who want Claude/Cursor-style agents to recall patterns, ADRs, and conventions via MCP without building custom vector storage first.
Skip if: One-off scripts with no repo, teams that already mandate a single enterprise knowledge base with strict ACLs, or builders who cannot run Node 18+ locally.
When should I use this skill?
When you need a persistent, searchable memory bank for AI agents that syncs with project documentation via MCP.
What do I get? / Deliverables
A running MCP memory bank tied to your workspace lets agents search and write durable memories that stay aligned with project documentation as you build and maintain the product.
- Compiled MCP server running for a project_id and workspace
- Searchable memory entries retrievable via memory_search
- New patterns and decisions stored via memory_write
Recommended Skills
Journey fit
Spans multiple journey phases - primary shelf plus alternate fits below.
Build → agent-tooling is the canonical shelf because setup centers on cloning agentMemory, compiling the MCP server, and wiring tools into the agent workspace. Agent-tooling captures MCP server lifecycle (install, compile, start-server with project_id and workspace path) rather than a single app feature integration.
Where it fits
Clone agentMemory, compile, and start-server so every feature branch chat can memory_search authentication patterns.
Record integration decisions with memory_write after wiring a third-party API so later agent tasks reuse the same key names and error handling.
Search prior incident decisions before patching a regression the agent helped ship months ago.
Retrieve documented tone and positioning patterns when drafting lifecycle emails or changelog copy.
How it compares
Skill-packaged MCP memory server for project-scoped recall—not a hosted analytics warehouse or a generic note-taking MCP with no doc sync story.
Common Questions / FAQ
Who is agent-memory-mcp for?
Solo and indie builders using AI coding agents who need persistent, typed memories (architecture, patterns, decisions) searchable from the agent via MCP.
When should I use agent-memory-mcp?
Use it during Build when wiring agent tooling, during Operate iterate when fixing production issues that depend on past decisions, and during Grow content when reusing documented patterns across features.
Is agent-memory-mcp safe to install?
It clones a third-party GitHub repo and runs a local Node MCP server with filesystem access to your chosen workspace; review the Security Audits panel on this page before granting agent tool access.
SKILL.md
READMESKILL.md - Agent Memory Mcp
# Agent Memory Skill This skill provides a persistent, searchable memory bank that automatically syncs with project documentation. It runs as an MCP server to allow reading/writing/searching of long-term memories. ## Prerequisites - Node.js (v18+) ## Setup 1. **Clone the Repository**: Clone the `agentMemory` project into your agent's workspace or a parallel directory: ```bash git clone https://github.com/webzler/agentMemory.git .agent/skills/agent-memory ``` 2. **Install Dependencies**: ```bash cd .agent/skills/agent-memory npm install npm run compile ``` 3. **Start the MCP Server**: Use the helper script to activate the memory bank for your current project: ```bash npm run start-server <project_id> <absolute_path_to_target_workspace> ``` _Example for current directory:_ ```bash npm run start-server my-project $(pwd) ``` ## Capabilities (MCP Tools) ### `memory_search` Search for memories by query, type, or tags. - **Args**: `query` (string), `type?` (string), `tags?` (string[]) - **Usage**: "Find all authentication patterns" -> `memory_search({ query: "authentication", type: "pattern" })` ### `memory_write` Record new knowledge or decisions. - **Args**: `key` (string), `type` (string), `content` (string), `tags?` (string[]) - **Usage**: "Save this architecture decision" -> `memory_write({ key: "auth-v1", type: "decision", content: "..." })` ### `memory_read` Retrieve specific memory content by key. - **Args**: `key` (string) - **Usage**: "Get the auth design" -> `memory_read({ key: "auth-v1" })` ### `memory_stats` View analytics on memory usage. - **Usage**: "Show memory statistics" -> `memory_stats({})` ## Dashboard This skill includes a standalone dashboard to visualize memory usage. ```bash npm run start-dashboard <absolute_path_to_target_workspace> ``` Access at: `http://localhost:3333` ## When to Use This skill is applicable to execute the workflow or actions described in the overview. ## Limitations - Use this skill only when the task clearly matches the scope described above. - Do not treat the output as a substitute for environment-specific validation, testing, or expert review. - Stop and ask for clarification if required inputs, permissions, safety boundaries, or success criteria are missing.