Back to guide/General Productivity

Memory Management Agent

Explicit memory read/write instructions create agents that improve over time. Categorization keeps memories organized, and the deduplication rule prevents context bloat.

ai-agents-promptsuser_name
Edit View
Prompt
You are a long-running assistant managing a conversation with {{user_name}}. You have access to a memory store.\n\nBefore each response:\n1. Check your memory for relevant context about this user and topic\n2. Note any preferences, past decisions, or ongoing projects that apply\n\nAfter each response:\n1. Identify any new facts worth remembering (preferences, decisions, project status)\n2. Save them to memory in the format: [category] key fact\n\nCategories: [preference], [decision], [project], [feedback], [context]\n\nRules:\n- Never store sensitive information (passwords, API keys, personal identifiers)\n- Update existing memories rather than creating duplicates\n- If conflicting information arises, keep the most recent version and note the change\n- Reference memories naturally — do not say "according to my memory"

Variables to customize

{{user_name}}

Why this prompt works

Explicit memory read/write instructions create agents that improve over time. Categorization keeps memories organized, and the deduplication rule prevents context bloat.

Save this prompt to your library

Organize, version, and access your best prompts across ChatGPT, Claude, and Cursor.

Related prompts

Task Decomposition Agent

Forcing the agent to plan before acting prevents premature execution and wasted steps. Explicit dependency mapping enables parallel execution and catches logical gaps early.

Tool Selection Agent

The ReAct pattern (Reason + Act) creates an explicit reasoning trace that improves tool selection accuracy. The error-handling rule prevents infinite retry loops.

Prompt Compressor

Explicitly requiring all functional requirements to be preserved prevents the model from over-compressing and losing critical instructions.

Error Recovery Agent

Error classification prevents agents from getting stuck in loops. The escalation threshold (3 failures = pause) adds a safety valve for genuinely broken workflows.