Agent Loop MCP
- Updated July 18, 2026
- meharajM/agent-loop-mcp
Agent Loop MCP is a MCP server that manages agentic memory for long-running agent loops over stdio.
About
Agent Loop MCP is a stdio Model Context Protocol server packaged on npm as @mhrj/mcp-agent-loop. It is positioned as an agentic memory manager, meaning it gives Claude Code, Cursor, or Codex a structured place to store and retrieve context across repeated tool loops instead of relying only on the chat window. developers running multi-day features or autonomous-ish task chains benefit when decisions, partial results, and session state can be rehydrated without re-pasting specs. Install aligns with other meharajM tooling: add the MCP entry pointing at the npm binary and optionally pull the companion skill via npx skills add. Complexity is intermediate because you must understand what the agent is allowed to remember and how that affects privacy. It complements planning skills but does not replace version-controlled docs or issue trackers.
- NPM package @mhrj/mcp-agent-loop (stdio transport, version 1.0.3)
- Described as agentic memory manager for long-running agent loops
- Optional install path: npx skills add meharajM/agent-loop-mcp
- GitHub: meharajM/agent-loop-mcp
Agent Loop MCP by the numbers
- Data as of Aug 10, 2026 (Skillselion catalog sync)
claude mcp add agent-loop-mcp -- npx -y @mhrj/mcp-agent-loopAdd your badge
Show developers this MCP server is listed on Skillselion. Paste this into your README.
| Package | @mhrj/mcp-agent-loop |
|---|---|
| Transport | STDIO |
| Auth | None |
| Last updated | July 18, 2026 |
| Repository | meharajM/agent-loop-mcp ↗ |
What it does
Run stdio MCP agent-loop locally so coding agents persist and recall loop-oriented memory across long build and operate sessions.
Who is it for?
Best when you're running extended Claude Code or Cursor sessions on one codebase and want MCP-backed memory.
Skip if: Skip if you already enforce all state in Notion or Git and want zero agent-side persistence.
What you get
Agents read and write loop memory through MCP so you iterate features and fixes with less re-explaining.
- Running stdio MCP memory manager for agent loops
- Cross-session recall for agent-driven build and iterate work
- Optional skills.sh skill install for documented workflows
By the numbers
- Package version 1.0.3
- Registry identifier: @mhrj/mcp-agent-loop
- Transport: stdio
README.md
Agentic Loop Memory Server ♾️
The industry-standard persistent memory and state manager for long-running agentic workflows.
Enable any AI model—especially smaller ones with limited context windows—to function with the persistence of high-end models. This project works as a two-part ecosystem: an MCP Server for state management and an Agent Skill for orchestration.
🛠 Complete Setup (Required)
For the best experience, you must install both the orchestration skill and the MCP server.
1. Install the Skill
Install instructions into your AI agent (Claude Code, Cursor, etc.):
npx skills add meharajM/agent-loop-mcp --yes
2. Configure the MCP Server
Add the following to your `mcp_config.json`:
{
"mcpServers": {
"agent-loop": {
"command": "npx",
"args": ["-y", "@mhrj/mcp-agent-loop"]
}
}
}
🌟 Why this approach is unique
Unlike passive memory tools, this is an Active State Manager. It monitors word counts to trigger compaction cycles and enforces a "Self-Healing Strategy" on every failure, preventing AI agents from getting stuck in mindless loops.
📂 Project Structure
src/: TypeScript source for the MCP server.skills/agentic-loop/SKILL.md: The instruction manual for the AI.build/: JavaScript artifacts.
📄 License
ISC
Recommended MCP Servers
0Latency Memory MCPPersistent memory layer for AI agents.
0nMCP — Universal AI API OrchestratorUniversal AI API Orchestrator — 1,554 tools, 96 services.
0xHumans Protocol MCPMCP for AI agents: financing, skills, lending on Base1k Patient MCP1k patient MCP server
1trippulse MCP1trip PULSE: 21-tool AI travel planner.
3D AI Agent Avatar MCP3D AI Agent Avatar — render any GLB, give it a Solana wallet, a voice, and pump.fun powers.90How it compares
Persistent agent-memory MCP, not a one-shot codegen skill or WhatsApp human-in-the-loop bridge.
FAQ
Who is agent-loop-mcp for?
Developers using MCP-capable agents who need durable loop memory across many tool calls on the same project.
When should I use agent-loop-mcp?
When builds or operate iterations span multiple sessions and you want the agent to recall prior loop state via MCP.
How do I add agent-loop-mcp to my agent?
Install @mhrj/mcp-agent-loop from npm, configure stdio MCP in your client, and optionally run npx skills add meharajM/agent-loop-mcp for the companion skill.