agentmemory-mcp-tools skill for Claude Code: map of agentmemory tools

Gives Claude a map of agentmemory's memory tools, what each does and what it needs, for people who run agentmemory and want the right tool called.

Installs: 11.2K; on skills.sh
Stars: 30.1K; on GitHub
Files: 2; besides SKILL.md
Updated: Oct 2026; last repo commit

agentmemory-mcp-tools is a reference skill. When Claude has to choose a memory tool, or a tool name or argument is unclear, it uses this skill to pick the narrowest tool and look up the exact fields in REFERENCE.md. The SKILL.md groups the tools into six families and starts with a save-then-recall quick start.

The SKILL.md sets user-invocable: false, so it is not a slash command. It bundles no scripts. It is an index of the tools, so it needs the agentmemory server running and its MCP tools connected.

Good for

People using agentmemory across a coding agent who want accurate tool calls
Developers checking what agentmemory can do over MCP
Anyone debugging a wrong tool name or argument

Install agentmemory-mcp-tools

Claude Code

Copy this prompt.

Install the agentmemory-mcp-tools 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 maps tools that the agentmemory server provides, so it needs the server and the MCP connection. The plugin route is the one the README recommends for Claude Code. The skills-only route installs all 17 skills and does not connect the MCP tools by itself. Pick one route, not both.

From the project README. Source.

What's in the skill

When it loads

Map of every agentmemory MCP tool, what each does, and its parameters. Use when choosing which memory tool to call, when a tool name or argument is unclear, or when answering what agentmemory can do via MCP.

Sections in SKILL.md

Quick start
Tool families
Workflow
See also
Reference

Key facts

Author: Rohit Ghumare (rohitg00)
Repository: rohitg00/agentmemory
Installs on skills.sh: 11,183 as of 2026-11-02
GitHub stars: 30,069 for the whole rohitg00/agentmemory repo
License: Apache-2.0 (plugin.json)
Last repo commit: 2026-10-29
What is inside: 3 files including SKILL.md: SKILL.md, REFERENCE.md and TROUBLESHOOTING.md. It bundles no scripts.
Tools it maps: 54 MCP tools, 8 of them in the lean core set (REFERENCE.md)
Part of: The agentmemory plugin (17 skills, 12 hooks, 54 MCP tools per plugin.json)
Needs: The agentmemory server running and its MCP tools connected

What it includes

memory_save, then memory_smart_search with targetLayer "memory"
memory_save takes content, concepts and files. memory_smart_search takes a query and a limit, and without targetLayer it also searches captured observations.
Try: “Save this decision to agentmemory, then find it again.”
Write · Quick start: save then recall
memory_save and memory_compress_file
REFERENCE.md says memory_compress_file shrinks a markdown file and creates a .original.md backup before writing. Automatic capture comes from hooks, not an MCP tool.
Write · Capture tools
Smart search, recall, file history, timeline and image search
Read · Retrieve tools
memory_sessions, memory_commits, memory_commit_lookup
They list sessions and look up which session produced a git commit.
Read · Sessions and commits
Lessons, graph queries, relations, patterns and crystallize
The family has memory_lesson_save, memory_lesson_recall, memory_graph_query, memory_relations, memory_patterns and memory_crystallize.
Write · Knowledge and graph
Create, append, get, list, replace and delete size-limited slots
The family has 6 tools, from memory_slot_create to memory_slot_delete.
Delete · Structured slots
memory_governance_delete, memory_audit, memory_verify, memory_heal, memory_diagnose
memory_governance_delete deletes chosen memories and leaves an audit trail. memory_heal fixes issues that diagnostics find.
Delete · Governance and health
Full table of 54 tools with parameters, required marks and the 8 core tools
Core tools are memory_consolidate, memory_diagnose, memory_lesson_save, memory_recall, memory_reflect, memory_save, memory_sessions and memory_smart_search.
Read · REFERENCE.md
Narrowest tool first, exact parameters, only documented fields
It tells Claude to prefer memory_smart_search for open recall and to separate matched_count 0 from excluded_by_budget.
Workflow

From the official docs

Permissions

Medium risk

The skill is only a map, but the tools it points to can delete memories, write files and share data. Ask Claude to confirm before it calls a delete or share tool.

What it can do

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

Watch out for

The tools it maps can delete memories
REFERENCE.md lists memory_governance_delete, memory_lesson_delete and memory_slot_delete. Ask Claude to confirm what it will remove before it calls them. github.com
Some tools write files or share memories
memory_compress_file, memory_obsidian_export and memory_claude_bridge_sync touch files. memory_team_share and memory_mesh_sync send memories to others, so check the target first. github.com
The plugin records your sessions through hooks
The README lists captured data: user prompts (privacy-filtered), tool input and output, and the assistant reply of each turn. Review what is stored in the viewer, and use the forget skill to remove entries. github.com

Technical details

Turning on read-only	No read-only switch. Set AGENTMEMORY_TOOLS=core on the daemon to load only the 8 core tools, then restart it and reconnect MCP. That set still includes memory_save and memory_lesson_save, so it is not read-only.
Talks to	The agentmemory server on localhost:3111. Peer agentmemory instances, when memory_mesh_sync is used
Scopes	The skill only indexes tools, but the tools it points to can save, compress, share and delete memories memory_compress_file rewrites a markdown file and keeps a .original.md backup memory_obsidian_export and memory_claude_bridge_sync write files, and memory_mesh_sync and memory_team_share send memories to peers or team members The plugin registers 12 hooks that run node scripts on events such as every prompt and every tool call The plugin ships a .mcp.json that starts a node bridge to the local server

Questions

What does the agentmemory-mcp-tools skill do?

It gives Claude a map of agentmemory's MCP tools, with what each does and its parameters. Claude uses it to pick the narrowest tool and to get argument names right.

How many tools does agentmemory have?

REFERENCE.md says 54 MCP tools, of which 8 are in the lean core set. The core set loads with AGENTMEMORY_TOOLS=core and the full set is the default.

How do I install agentmemory-mcp-tools?

Use the install box at the top of this page. Start the agentmemory server first, then add the plugin, because the tools come from the server. Installing the skill folder alone does not give you the tools.

Why do I only see a few memory tools?

Missing tools can mean a stopped MCP process, a disconnected daemon, or a reduced tool set. SKILL.md says a native memory provider may expose fewer tools, and Hermes, OpenClaw, Pi and OMP need their own MCP connection for the full set. The troubleshooting file walks through the checks.

How do I save and find a memory?

Call memory_save with the content, concepts and files, then memory_smart_search with a query. Set targetLayer to "memory" to search only saved knowledge. The remember and recall skills wrap these two calls.

Can the tools delete my memories?

Yes. REFERENCE.md lists memory_governance_delete, memory_lesson_delete and memory_slot_delete. Ask Claude to confirm what it will delete before it calls them.

Is it safe to use?

The skill is a reference file and bundles no scripts. The risk is in the tools and plugin around it: they can delete memories and write files, and the plugin records sessions through hooks. See the safety section before you install.

Information

updated Nov 2, 2026

Repository: rohitg00/agentmemory
Folder: plugin/skills/agentmemory-mcp-tools

rohitg00/agentmemory skills.sh

Sources

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

Show all 7

agentmemory README · github.com
agentmemory-mcp-tools SKILL.md · github.com
agentmemory-mcp-tools REFERENCE.md · github.com
agentmemory plugin.json · github.com
agentmemory-mcp-tools TROUBLESHOOTING.md · github.com
agentmemory plugin .mcp.json · github.com
agentmemory plugin hooks.json · github.com