agentmemory-rest-api Skill (agentmemory) for Claude: What It Does and How to Install It

Teaches Claude to call the agentmemory memory server over HTTP, for hosts where its MCP tools are missing.

Installs: 11.1K; on skills.sh
Stars: 30.1K; on GitHub
Files: 2; besides SKILL.md
Updated: Oct 2026; last repo commit

This is a reference skill for agentmemory, a memory server for coding agents. It says REST is agentmemory's primary surface and MCP is a bridge on top of it, with every memory operation available under http://localhost:3111/agentmemory/*. The quick start checks that the server is alive, saves a memory and searches for one.

Claude loads it on its own, because the file sets user-invocable: false. That happens when it calls agentmemory over HTTP, when MCP is not available or when a host does not speak MCP. REFERENCE.md lists 146 endpoints. It needs the agentmemory server running, and protected endpoints need a bearer secret.

Good for

Developers wiring agentmemory into a host that has no MCP support
Anyone whose agentmemory MCP tools are missing and who needs a fallback
People who want to save or search memories with curl

Install agentmemory-rest-api

Claude Code

Copy this prompt.

Install the agentmemory-rest-api skill in Claude Code. Run this command:

npx -y @agentmemory/agentmemory@latest

Then check that it loaded and tell me how to use it.

2
Paste it into Claude Code. It shows the change and asks before running it.

Requires Node.js 20 or newer.

This skill only works when the memory server is running. It calls the server's HTTP API, and the file says to use it when the MCP tools are unavailable. The plugin route is the one the README recommends for Claude Code. The skills-only route installs all 17 skills but does not start the server. Pick one route, not both.

From the project README. Source.

What's in the skill

When it loads

The agentmemory HTTP REST API surface, the primary protocol for talking to the memory server. Use when calling agentmemory over HTTP, when MCP is unavailable and you need a fallback, or when integrating a host that does not speak MCP.

Sections in SKILL.md

Quick start
Auth
Conventions
See also
Reference

Key facts

Author: Rohit Ghumare (rohitg00), per plugin.json
License: Apache-2.0 (repo and plugin.json)
Installs on skills.sh: 11,116 (Nov 2, 2026)
GitHub stars: 30,069 for the whole rohitg00/agentmemory repo (Nov 2, 2026)
Last repo commit: Oct 29, 2026
What it includes: 3 files including SKILL.md: REFERENCE.md (endpoint list) and TROUBLESHOOTING.md (no scripts)
Part of: The agentmemory plugin (17 skills, 12 hooks, 54 MCP tools per plugin.json)
Needs: The agentmemory server running (default http://localhost:3111) and a bearer secret from AGENTMEMORY_SECRET or ~/.agentmemory/secret

What it includes

GET /agentmemory/livez, which is public
A curl call that checks the server is up. SKILL.md says liveness is public.
Try: “Check that the agentmemory server is running.”
Network · Liveness check
POST /agentmemory/remember with content and concepts
The quick start saves 'chose JWT refresh rotation' with the concept jwt-refresh-rotation. A save returns 201.
Write · Save a memory
POST /agentmemory/smart-search with a query and limit
The quick start searches for 'auth token strategy' with a limit of 5. Reads return 200.
Network · Search memories
Authorization header from AGENTMEMORY_SECRET or ~/.agentmemory/secret
Protected endpoints need a bearer secret, including on localhost. The example loads it without printing it, for the fixed loopback URL only.
Files · Bearer authentication
Status codes, documented fields only, ports
Validation errors return 400. REST defaults to port 3111, the viewer to 3113 and the internal engine WebSocket to 49134.
Read · Conventions
146 registered endpoints with their HTTP methods
A table generated from the server's source. It includes forget, bulk delete and graph reset endpoints.
Read · REFERENCE.md endpoint list
Which REST call replaces each memory skill (TROUBLESHOOTING.md)
For example recall maps to POST /agentmemory/smart-search and commit-history to GET /agentmemory/commits.
Read · REST fallback map

From the official docs

Permissions

Medium risk

The main risk is that Claude can reach every memory endpoint, including deletes, with your local secret, and the plugin records your sessions. Ask for confirmation before deletes and review what the viewer shows is stored.

What it can do

Read your data
Create and change things
Delete things
Read-only mode

Watch out for

The endpoint list includes deletes and resets
REFERENCE.md lists POST /agentmemory/forget, POST /agentmemory/governance/bulk-delete, DELETE /agentmemory/governance/memories, POST /agentmemory/evict, POST /agentmemory/auto-forget and POST /agentmemory/graph/reset. The quick start uses only remember and smart-search. Ask Claude to confirm before any call that removes data. github.com
The plugin records your sessions through hooks
The README says the Claude Code hooks capture user prompts (privacy-filtered), tool use and results, errors and the assistant's reply for each turn. Review what is stored in the viewer at localhost:3113 before you rely on it. github.com
Setup runs npm code and starts a local server
The README's setup command is npx -y @agentmemory/agentmemory@latest, which pulls the latest package and starts the memory server. The plugin's .mcp.json also starts a node bridge script from the plugin folder. github.com
It reads a local secret file
The quick start falls back to $HOME/.agentmemory/secret and says the command is for the fixed loopback URL only. TROUBLESHOOTING.md says never to send a local-file secret to a remote URL and to reject redirects on authenticated requests. github.com

Security advisories

No published security advisories as of Nov 2, 2026.

Technical details

Turning on read-only	No read-only switch is documented for these skills.
Talks to	Requests to the memory server, which the README says binds to loopback (127.0.0.1) by default. A remote server if you set AGENTMEMORY_URL; TROUBLESHOOTING.md requires HTTPS for authenticated requests to non-loopback URLs. The hook scripts call to AGENTMEMORY_URL, which defaults to http://localhost:3111; if you point it at a remote server, hook data goes there
Sign-in and keys	A local server makes a random secret at ~/.agentmemory/secret (mode 0600, per the README) and bundled clients read it. A remote server needs AGENTMEMORY_SECRET in your environment.
Scopes	Runs curl against the memory server at http://localhost:3111, such as saving a memory and searching Reads the bearer secret from AGENTMEMORY_SECRET or from ~/.agentmemory/secret and sends it in an Authorization header REFERENCE.md lists 146 endpoints, including forget, bulk delete, evict, auto-forget and graph reset The Claude Code plugin registers hooks for 12 events (such as every prompt, every tool call and session end), each running a node script from the plugin The plugin ships a .mcp.json that starts a node bridge (plugin-bridge.mjs) to the memory server Setup runs npx -y @agentmemory/agentmemory@latest, which starts a server on localhost:3111 with a viewer on localhost:3113 In the plugin source, session-start.mjs and pre-tool-use.mjs read AGENTMEMORY_URL and the bearer secret (AGENTMEMORY_SECRET, an .env file or ~/.agentmemory/secret) and call the memory server; session-start.mjs and stop.mjs also run git rev-parse

Questions

What does the agentmemory-rest-api skill do?

It tells Claude how to talk to the agentmemory memory server over plain HTTP. Claude uses it to check the server, save a memory, search memories and find the right endpoint.

How do I install it?

Use the install box at the top of this page. Start the agentmemory server first, then add the plugin or the skills. The skill calls the server, so the folder alone does nothing.

When would I use REST instead of MCP?

SKILL.md says to use it when calling agentmemory over HTTP, when MCP is unavailable and you need a fallback, or when a host does not speak MCP.

Do I need a secret to call the API?

Yes for protected endpoints, even on localhost. The skill reads AGENTMEMORY_SECRET or the generated file at ~/.agentmemory/secret. The liveness check is public.

Can it delete memories?

The quick start only saves and searches. REFERENCE.md also lists endpoints for forget, bulk delete, evict, auto-forget and graph reset, so Claude can call them if you ask or allow it. Tell it to confirm before any call that deletes.

Is it safe?

The skill itself is text and bundles no scripts. It does tell Claude to run curl against your local memory server with a bearer secret read from your machine. The setup around it is larger: the plugin adds hooks that record your sessions and starts a local server. See the safety section on this page.

Information

updated Nov 2, 2026

Repository: rohitg00/agentmemory
Folder: plugin/skills/agentmemory-rest-api

rohitg00/agentmemory skills.sh

Sources

Docs, issue trackers and discussions this page is built from.

Show all 9

agentmemory README · github.com
skills CLI commands · github.com
agentmemory-rest-api SKILL.md · github.com
plugin.json · github.com
REFERENCE.md · github.com
TROUBLESHOOTING.md · github.com
plugin hooks.json · github.com
plugin .mcp.json · github.com
pre-tool-use.mjs · github.com