More screenshots

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    second brain for permanent memory to never forget anything

    2

    A skill that makes your agents never forget a single thing. This is the first step towards having your own jarvis. You will never need to describe anything to any agent across any chat in Claude code.

    $4.99

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned
    second brain for permanent memory to never forget anything

    second brain for permanent memory to never forget anything

    Example session with this skill installed

    file this chat and record that we decided to use Tailwind instead of Bootstrap because of the smaller bundle size requirements for the mobile version.

    • Read your context and instructions
    • Compiled the second brain

    The brain already held a September decision to use Bootstrap. Here is what the agent did with the new one.

    Filed this chat as chat-2026-10-17-tailwind-over-bootstrap.

    ActionItemDetail
    New decisionengineering--css-framework-tailwindUse Tailwind instead of Bootstrap: the mobile version has smaller bundle size requirements.
    Supersededengineering--css-framework-bootstrapMarked superseded and linked to the new note. Body kept, so the history of why stays readable.
    IndexINDEX.mdTailwind is listed as the active decision; Bootstrap moved under Superseded.
    Chat cardchats/chat-2026-10-17-tailwind-over-bootstrap.mdSummary, decisions and links.
    BriefBRIEF.mdNo change needed.

    No lessons or handoffs. If Bootstrap was already built into screens and is now being ripped out, say so and I'll add a lesson.

    The new note

    ---
    id: engineering--css-framework-tailwind
    title: Use Tailwind for styling
    type: decision
    area: engineering
    status: active
    created: 2026-10-17
    source: [chat-2026-10-17-tailwind-over-bootstrap]
    supersedes: [engineering--css-framework-bootstrap]
    superseded_by: null
    tags: [frontend, css, mobile]
    ---
    Use Tailwind for styling instead of Bootstrap. Changed from [[engineering--css-framework-bootstrap]].
    **Why:** the mobile version has smaller bundle size requirements, and Tailwind ships only the CSS that is used.
    **How to apply:** new UI, and the mobile version in particular, is styled with Tailwind; don't add Bootstrap to new code.
    

    The old note, now superseded (not deleted)

    id: engineering--css-framework-bootstrap
    title: Use Bootstrap for styling
    status: superseded
    superseded_by: engineering--css-framework-tailwind
    

    The index

    
    ## engineering
    ### decision
    - [Use Tailwind for styling](areas/engineering/notes/engineering--css-framework-tailwind.md) — Use Tailwind for styling instead of Bootstrap…
    
    ## Superseded
    - [Use Bootstrap for styling](areas/engineering/notes/engineering--css-framework-bootstrap.md) — All UI uses Bootstrap 5. (superseded by engineering--css-framework-tailwind)
    

    Next time anyone asks "what CSS framework do we use, and why?", the agent answers Tailwind, because of the mobile bundle size, and it replaced Bootstrap on 2026-10-17, with the note ids as sources. No more re-explaining.

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

    About this skill

    Stop repeating context to your AI agent every time you start a new chat. This skill implements a durable, file-based memory system that lives directly in your project folder as a collection of plain Markdown files. It ensures your agent remembers past technical decisions, architectural preferences, and open questions across sessions without requiring a database or external API.

    What it does

    • Persistent Memory creates a structured folder hierarchy to store project facts, decisions, and lessons.
    • Automated Filing processes finished chats into concise notes, chat cards, and cross-referenced indexes.
    • Context Recovery imports historical chat logs from Claude Code and Codex to prepopulate your project brain.
    • Conflict Resolution manages superseded decisions by linking old notes to new ones, maintaining a clear audit trail.
    • Inbox Management facilitates handoffs between different project areas like engineering, product, and marketing.

    How it works

    1. Setup initializes the folder structure, a high-level project brief, and hooks for your agent's system instructions.
    2. File extracts settled facts from the current session and updates the INDEX.md and area-specific notes.
    3. Recall searches the brain to answer questions about past decisions with citations to specific chat cards.
    4. Weekly triggers a review to merge duplicates, clear inboxes, and update project priorities in the brief.

    Frameworks & tools

    This skill works with any agent that has file-system access, including Claude Code, Cursor, and GitHub Copilot. It requires no external dependencies, relying entirely on Markdown and YAML front matter.

    Why this beats prompting it yourself

    Manual prompts for "remembering" things often result in messy, redundant, or hallucinated history. This skill enforces a strict "one fact per file" architecture and a superseding logic that prevents conflicting information from polluting your agent's context.

    Use cases

    • Maintaining Architectural Logic so the agent remembers why a specific library was ruled out weeks ago.
    • Onboarding Subagents by providing a single BRIEF.md that summarizes the entire project state.
    • Tracking User Preferences for coding style, naming conventions, and deployment workflows.

    Known limitations

    Does not support Cursor or Copilot chat history recovery due to their database-backed storage. Requires manual agent permission to read files outside the project root for recovery tasks.

    Filing happens when you ask (or accept the agent's suggestion at the end of a chat); nothing runs in the background. Recovering old chats files up to 10 per run and uses your agent's normal usage, so a long history takes several runs. Cannot recover history from Cursor or Copilot, which store chats in a database. Needs an agent that can read and write files; works best with agents that load project instructions (CLAUDE.md, AGENTS.md, Cursor rules).

    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

    Money-back guarantee

    Doesn't do what its page promises? You get your money back.

    Trust & safety

    Security scan

    Security scan passed 1 month ago

    Requires

    none.

    Permissions

    Read Files
    Write Files

    File Scopes

    brain/**
    CLAUDE.md
    AGENTS.md
    .cursor/rules/brain.mdc
    .github/copilot-instructions.md
    ~/.claude/projects/**
    ~/.codex/sessions/**

    Writes only markdown files inside the brain/ folder, plus one short "Memory" pointer in the agent's instruction file (CLAUDE.md, AGENTS.md, Cursor rules or Copilot instructions) so new chats read the brief. Reads ~/.claude/projects and ~/.codex/sessions only when the user asks to "recover my old chats", only after they confirm which chats, and never changes those files. No shell commands, no network access, no environment variables.

    Listed1 month ago

    What's inside

    Frequently Asked Questions

    Every decision, fact and preference gets its own short note with the reason behind it. Your agent reads a one-page brief at the start of each chat and checks the index before asking you to repeat yourself. When you change your mind, the old note is marked superseded instead of deleted, so the agent always knows what's current and why it changed.

    This skill is compatible with any AI agent that has file-system access, such as Claude Code, Cursor, and GitHub Copilot, as it relies entirely on Markdown and YAML.

    One SKILL.md that teaches your agent the whole method: setting up the brain, filing chats into notes, superseding old decisions, answering questions with sources, handoffs between areas, a daily brief, a weekly clean-up, and recovering past Claude Code and Codex chats. No scripts or apps; your agent does the work with its normal file tools.

    No external database is required; the skill creates a structured folder of Markdown files directly within your project directory for easy versioning and portability.

    While it can recover history from Claude Code and Codex, it currently cannot automatically import Cursor or Copilot chat history due to their proprietary database structures.

    About a minute. The agent asks two questions (where the brain lives and which areas you want), creates the folders and a short brief, and adds a three-line "Memory" pointer to your agent's instruction file so every new chat reads it.

    No. It only files what was actually settled in the chat. Ideas you didn't accept are saved as open questions, and it never changes your brief without asking.

    Yes. Everything is plain markdown in a folder you own. Edit, delete or move any note, and the agent picks up the change next time it reads the index.

    New in Agents & Orchestration