Back to Cookbook
OpenClaw recipe

Memory-Aware Research Agent

Research that builds on what you've already found

A research agent that tracks prior sessions and avoids redundant searches. It surfaces gaps in your existing knowledge base and picks up exactly where you left off.

CommunitySubmitted by CommunityWork5 min
Try in KiloClawFree 7-day trial

INTEGRATIONS NEEDED

PROMPT

Create a skill called "Memory-Aware Research Agent". Before researching any topic, read the session log to see what has already been covered and which sources have been reviewed. Focus exclusively on gaps and new angles. After each session, append a structured summary — key findings, sources consulted, open questions — to the log so future runs can build on it. Avoid re-summarising content already recorded. Deliver findings with named sources and working URLs.

How It Works

The Memory-Aware Research Agent keeps a running record of every topic,

source, and conclusion from past research sessions. Before diving into a

new query it checks what you already know, skips sources you've already

reviewed, and focuses effort on genuine gaps.

What Others Get

Research that compounds over time. Instead of re-covering the same ground,

each session builds on the last — so your knowledge base grows rather than

restarting from scratch every time.

Setup Steps

  1. Create a skill called "Memory-Aware Research Agent" using the prompt below
  2. Point the agent at a persistent store for session notes (a file, GitHub issue, or similar)
  3. Kick off your first research query — the agent logs what it finds
  4. On subsequent runs, pass the session log so the agent can resume intelligently
Tags:#research#memory#automation#productivity