remember

rohitg00/agentmemory

Save insights and decisions to searchable long-term memory with concept tags.

What is remember?

Preserves knowledge, decisions, and learnings in agentmemory's persistent storage with searchable concept tags. Use when the user explicitly asks to remember, save, note, or preserve information for future sessions.

Saves content to long-term memory with user's original phrasing preserved
Tags memories with 2-5 specific, lowercased concept phrases for reliable retrieval
Associates file paths with saved memories for context
Returns a memory ID and confirms the concepts used for tagging
Supports multi-agent setups via agentId parameter
Allows updating facts by saving corrected versions that supersede old records

How to install remember

npx skills add https://github.com/rohitg00/agentmemory --skill remember

Prerequisites

agentmemory installed and configured
recall skill available to retrieve saved memories

Claude Code

Cursor

Windsurf

Cline

How to use remember

1.Extract the core insight, decision, or fact from what the user wants to remember
2.Create 2-5 specific, lowercased concept phrases (e.g., 'jwt-refresh-rotation' not 'auth')
3.Identify any referenced file paths, absolute or repo-relative
4.Call memory_save with content, concepts (comma-separated), and files (comma-separated)
5.Confirm the save and echo back the exact concepts for future retrieval

Use cases

Good for

Preserve architectural decisions like token rotation patterns for future reference
Save API contract details, configuration rules, or deployment procedures
Store project-specific learnings or debugging insights discovered during development
Keep track of user preferences, constraints, or business rules mentioned in conversation
Document code patterns or solutions that should be reused across sessions

Who it's for

Developers building multi-session AI agents
Teams needing persistent knowledge across agent interactions
Anyone using agentmemory for long-term context retention

remember FAQ

What's the difference between remember and lesson?

remember stores facts and decisions; lesson stores behavioral rules learned from corrections.

How specific should concept tags be?

Use specific, retrievable terms like 'jwt-refresh-rotation' rather than generic ones like 'auth'. Aim for 2-5 concepts per memory.

Can I update a memory I saved?

Yes—save the corrected version outright. Near-duplicate content supersedes the old record while keeping it in the version chain.

Does this work in multi-agent setups?

Yes, pass agentId in the memory_save call to ensure the memory lands in the correct agent's scope.

What if I save something by mistake?

Use the forget skill to remove a memory you saved incorrectly.

Full instructions (SKILL.md)

Source of truth, from rohitg00/agentmemory.

name: remember description: Save an insight, decision, or learning to agentmemory's long-term storage with searchable concept tags. Use when the user says "remember this", "save this", "note that", "don't forget", or wants to preserve knowledge for future sessions. argument-hint: "[what to remember]" user-invocable: true

The user wants to save this to long-term memory: $ARGUMENTS

Quick start

memory_save {
  "content": "We rotate JWT refresh tokens on every use; the old token is revoked server-side in auth/refresh.ts.",
  "concepts": "jwt-refresh-rotation, token-revocation, auth-flow",
  "files": "src/auth/refresh.ts"
}

Expected output:

Saved memory abc12345 with 3 concepts: jwt-refresh-rotation, token-revocation, auth-flow.

Why

A memory is only as useful as the terms that retrieve it. Tag with specific concepts so a future recall finds it, and preserve the user's own phrasing.

Workflow

Pull the core insight, decision, or fact out of $ARGUMENTS.
Extract 2-5 lowercased concept phrases. Prefer specific over generic (jwt-refresh-rotation beats auth).
Extract referenced file paths (absolute or repo-relative). Empty if none.
Call memory_save with content, concepts (comma-separated string), and files (comma-separated string). In a multi-agent setup pass agentId so the memory lands in the right agent's scope.
Confirm the save and echo the concepts so the user knows the retrieval terms.
To update a fact, save the corrected version outright: near-duplicate content supersedes the old record, which leaves recall but stays in the version chain.