Agent Memory Scope and Retention Architect
Design what an AI agent may remember, for whom, for how long, with which evidence, controls, deletion paths, and privacy boundaries.
Agent Memory Scope and Retention Architect is a free AI skill for agent systems & llm workflows. Design what an AI agent may remember, for whom, for how long, with which evidence, controls, deletion paths, and privacy boundaries. It works with Claude, ChatGPT and is ready to use out of the box.
About this skill
Create a safe memory architecture for assistants and agents covering session memory, user preferences, project context, retrieved knowledge, summaries, tool state, long-term storage, access boundaries, correction, deletion, and evaluation.
What it does
It defines memory categories, write and read rules, consent, retention, provenance, conflicts, privacy controls, failure behavior, and tests.
What is included
- Memory category map
- Write and read policies
- User, workspace, and project boundaries
- Consent and visibility rules
- Retention and deletion
- Provenance and confidence
- Conflict and correction handling
- Security and privacy controls
- Evaluation metrics
- Implementation roadmap
How to use it
1. Download the agent-memory-scope-and-retention-architect-SKILL.md file 2. Upload it to Claude, ChatGPT, or your agent-design workspace 3. Provide the agent purpose, users, data categories, tools, storage, privacy requirements, and failure risks 4. Do not include production secrets or unnecessary personal data 5. Have privacy, security, product, and engineering owners approve long-term memory behavior
Examples
Design memory for a project-management AI assistant. It may remember user preferences, project decisions, owners, and deadlines but must not retain private chat content or cross workspace boundaries.
A memory taxonomy, read and write rules, consent, workspace boundaries, retention periods, provenance fields, correction and deletion flows, conflict handling, privacy controls, metrics, and tests.
FAQ
Should an agent remember everything?
What is the difference between session and long-term memory?
Can users correct remembered information?
How is memory provenance handled?
Can memory cross user or workspace boundaries?
How is this different from a vector database design?
Related Skills
Agent Memory and Context Governance Architect
Design agent memory governance with memory types, relevance, consent, retention, editing, deletion, isolation, retrieval, and safety.
AI Workflow Architect
Design reliable AI-agent and LLM workflows from a business goal, including steps, tools, prompts, controls, and implementation guidance.
AI Interface Interaction Designer
Design AI interfaces with context, streaming, citations, tool use, approvals, uncertainty, memory, errors, privacy, and user control.
Related Prompts
Agent Memory Architecture Planner
Design a practical memory system for an AI agent, including what to retain, how to retrieve it, and when to forget it.
Data Retention and Deletion Workflow
Plan an auditable automation that identifies eligible records, honors holds, verifies deletion across systems, and handles backups and failures.
Multi-Agent Role Designer
Designs a clean division of labor for a multi-agent system, defining each sub-agent's role, boundaries, and handoff protocol so agents don't overlap or conflict.