More screenshots

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    simem: Super Intelligence Memory

    1

    It doesn't know what you built yesterday, what you decided, or how to run the tests, so it rediscovers the project or you explain it again.

    Free

    Security scanned
    simem: Super Intelligence Memory

    simem: Super Intelligence Memory

    Example session with this skill installed

    What did we decide about the database last week? Also remember that integration tests must run against a real Postgres, never mocks.

    • Read your context and instructions
    • Compiled the simem

    Here is what earlier sessions saved about the database, and I've stored your new rule.

    What we decided

    I searched project memory with recall("database decision"):

    #MemoryCategorySaved
    1Use PostgreSQL 16 for the main app; SQLite only for local tooling.decisionOct 28
    2Migrations live in db/migrations and run with npm run db:migrate.factOct 28
    3Don't add an ORM; queries go through the thin db/query.ts wrapper.constraintOct 30

    The reason recorded with #1: the team needs JSONB and row-level security, which SQLite can't provide.

    Saved to memory

    memory_save
      text:       "Integration tests must run against a real Postgres database, never mocks."
      kind:       semantic
      category:   constraint
      importance: 8
    

    This note now comes back automatically in any future session that mentions tests or the database, in Claude Code, Codex or Cursor, under any account.

    Next steps

    • Ask "how much did simem save?" to see your token usage and savings for the last 7 days.
    • Say "show my memory" to open the private dashboard on 127.0.0.1, where you can search, edit or forget notes.
    • If a note goes stale, say "forget the ORM rule" and I'll remove it.

    Connects securely to your tools. The creator never sees your data.

    About this skill

    Every new agent session starts from scratch. It doesn't know what you built yesterday, what you decided, or how to run the tests, so it rediscovers the project or you explain it again. Switch from Claude Code to Codex, or to another account or worktree, and the memory is gone too.

    simem (Super Intelligence Memory) keeps that memory on your machine and gives it to every agent and every account you use: Claude Code, Codex and Cursor, in the desktop apps and the CLIs. This skill teaches your agent when and how to use it.

    What it does

    • Recall past decisions. Ask "what did we decide about payments?" and the agent searches everything earlier sessions saved, then opens the full note when it needs it.
    • Remember on request. Say "remember that…" and the agent saves it as one clear fact with a category (decision, constraint, preference, fact, todo) and an importance score.
    • Forget on request. Say "forget that…" and the agent finds the note and removes it.
    • Show token savings. Ask "how much did simem save?" and get your real token usage plus what simem kept out of the context, with measured and estimated numbers labeled.
    • Open the memory dashboard. A private page on 127.0.0.1 to search, edit or forget memories.
    • Set itself up. If simem isn't installed, the bundled setup guide installs the CLI and runs simem init, asking you before every step.

    Why it saves tokens

    Agents re-read the whole conversation on every turn, and much of it is long tool output. simem trims those reports before they go into the context, keeping errors, failures and summaries, and stores the full output locally in case the agent needs it. On real sessions it cut tool reports by 22% overall and test reports by 46%; full test runs and package installs shrink by 81 to 99%.

    Setup (handled by the skill)

    1. Install the CLI: npm install -g @simem/cli (Node.js 22.5+) or Homebrew (brew tap thinhledev/simem, brew trust --formula thinhledev/simem/simem, brew install simem).
    2. Run simem init once per machine. It connects simem to your agents, imports this project's past sessions, and downloads a small local embedding model (31 MB, MIT-licensed).
    3. Restart your agent app.

    Privacy

    Everything stays on your machine in ~/.simem. Nothing is uploaded and there is no simem server. Undo any time with simem init --remove.

    Open source (MIT): https://github.com/thinhledev/simem


    New to Agensi? Create your free account through my referral link to use this skill. Disclosure: I earn a commission on skills you buy in your first 6 months; it costs you nothing extra.

    Needs the simem CLI installed locally (Node.js 22.5+ or Homebrew); the skill guides setup. Memory lives on one machine by default; move it with simem memory export / import or point memory.syncDir at a folder you already sync. Cursor gets the memory tools but not the automatic hooks.

    How to use

    1. 1

      Connect your agent to the Agensi MCP

      Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup

    2. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 3

      Ask for what you need

      Your agent loads this skill and uses it on your task. It always asks before any payment.

    Reviews

    No reviews yet

    Be one of the first to try it. Every listed skill goes through the trust checks below.

    Security scanned

    Checked by our 8-point scan before listing

    Fresh listing

    Recently published to Agensi

    Free forever

    No account required to browse

    Trust & safety

    Security scan

    Security scan passed 11 days ago

    • Free with an Agensi account

    Requires

    simem CLI (@simem/cli)
    Node.js 22.5+ or Homebrew
    macOS or Linux
    Node.js
    macOS

    Permissions

    Write Files
    Network Access
    Terminal / Shell
    Read Files

    Shell: runs simem commands (recall, stats, dashboard, doctor) and, only after you approve, installs the CLI and runs simem init. Write files: simem init adds hook and MCP entries to your agent settings and stores memory in ~/.simem. Network: only for installing the CLI and the one-time 31 MB embedding model download. Nothing is uploaded.

    Listed11 days ago

    What's inside

    Frequently Asked Questions

    A context window is cleared when the session ends or compacts. simem keeps decisions, constraints and build/test commands on your machine and brings the relevant ones back in every new session, in any agent app and under any account.

    Claude Code and Codex (desktop apps and CLIs) get everything: hooks, memory tools and this skill. Cursor gets the memory tools. Any agent that can run shell commands can use the simem CLI fallback.

    The simem skill (recall, remember, forget, token stats, dashboard, doctor) plus a setup guide that installs the open-source simem CLI and runs simem init, asking before each step. simem itself is free and MIT-licensed.

    Yes. Memory stays local by default, but you can move it with simem memory export / import, or point memory.syncDir at a folder you already sync (iCloud, Dropbox), optionally encrypted.

    Yes. Ask how much simem saved to get your real token usage and savings (measured and estimated numbers are labeled), and simem doctor checks the setup and tells you what to fix.

    New in LLM Ops