memory

Retrieve and persist cross-task knowledge using the memory MCP tool.

Updated Aug 23, 2026
One-click install
npx skills add https://github.com/SalesTeamToolbox/frood --skill memory-salesteamtoolbox
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: memory
Source: https://github.com/SalesTeamToolbox/frood/tree/main/skills/builtins/memory
Command: npx skills add https://github.com/SalesTeamToolbox/frood --skill memory-salesteamtoolbox

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill prevents agents from repeatedly starting over by enabling reliable cross-task, cross-session recall of user preferences, decisions, and prior work.

Core Features & Use Cases

  • Persistent long-term memory: store consolidated facts and preferences into MEMORY.md for durable guidance across tasks.
  • Event logging and traceability: record significant outcomes and decisions into HISTORY.md to preserve an audit trail of what happened and why.
  • Semantic recall with vector search: retrieve relevant past conversations and knowledge using Qdrant when exact keywords are unavailable.
  • Fast session caching: use Redis to speed up active-session operations like embedding lookups with TTL-based expiry.
  • Correct-by-validation memory ops: support forget, correct, and strengthen to reduce hallucinated or outdated memories and improve confidence over time.

Quick Start

Use the memory tool to search for relevant prior context before starting a new task by telling the agent: "Search Frood memory for what we decided previously that could help with this change."

Frequently Asked Questions about memory

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I enable persistent cross-session context for AI agent orchestration?

Store consolidated facts and preferences into a durable file like MEMORY.md, and record significant outcomes into a history log to preserve a reliable audit trail of agent decisions across tasks.

How does semantic search work for recalling prior agent conversations without exact keywords?

It also uses Redis for fast session caching to speed up active-session operations like embedding lookups with TTL-based expiry, ensuring quick retrieval during ongoing tasks.

Can I correct or forget outdated memories to prevent AI agent hallucination?

It applies to starting new tasks, responding to questions about what was remembered, and grounding decisions in prior conversations and stored learnings.

What is the best way to search stored context before starting a new task with an agent?

This ensures the agent grounds its current task in relevant prior context, preventing it from repeatedly starting over and making the task smarter with retained knowledge.

Does cross-task memory work with Redis and Qdrant for agent knowledge retrieval?

This combination ensures both durable long-term recall across sessions and fast performance during active tasks, preventing agents from starting over.