Skip to content

Past Session Memory Search

Lets the agent search claude-mem's saved memory of earlier sessions before fetching full details.

Stars
110,174
Licence
Apache-2.0
Files
1 file

What it does

Once installed, the agent can look back at a persistent memory database that claude-mem keeps across past coding sessions. Instead of guessing, it runs a three-step process: search for a short index of matching notes, look at the timeline around promising entries, then fetch full details only for the entries that matter. This helps it answer questions like whether a bug was already fixed or how a problem was handled last time.

Great for

  • Developers who use claude-mem and want the agent to recall earlier fixes and decisions
  • Teams reviewing how a problem was solved in previous sessions
  • Anyone who wants to cut token use by filtering memory before loading full records

Skip it if

  • Questions about the current conversation, since the skill is meant for previous sessions only
  • Setups without the claude-mem memory database or its MCP tools, because the skill has nothing to query
  • Situations requiring authoritative records: observations are summaries and the raw tool-output layer can contain large stored payloads that may include sensitive data

Try asking

  • “Did we already fix this authentication token bug?”
  • “How did we solve the migration problem last time?”
  • “What happened in the project last week?”
  • “Recover the exact output of the migration command we ran last week”
Step-by-step guide: how to use Past Session Memory SearchInstall, first run, examples, tips and FAQs

Install in 3 steps

  1. 1

    Download the skill

    The ZIP contains the 1 original files, the Apache-2.0 licence and an attribution note.

    mem-search.zip
  2. 2

    Add it to your agent

    Claude Code (all projects)

    mkdir -p ~/.claude/skills && unzip mem-search.zip -d ~/.claude/skills/

    Claude Code / Agent SDK (this project only)

    mkdir -p .claude/skills && unzip mem-search.zip -d .claude/skills/

    Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.

  3. 3

    Ask for the task

    No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.

Get better results

  • Start with a specific search query and a project filter to keep results small
  • Only ask for raw tool output when the summarized observations do not contain the needed detail
  • If you use a self-hosted or CMEM Pro connection, check the advertised schema since parameter names differ

What's inside

  • SKILL.md

Redistributed unmodified from thedotmack/claude-mem under the Apache-2.0 licence. The assessment was drafted with AI from the skill's files on 2026-10-29. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.