Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIDeepSeekDeepSeekMuseMuse+22 more

    Persistent Agent Memory

    1

    A buyer-owned workflow for the forgetting problem: agents say sorry for the oversight and then forget again, because the session ends and nothing durable is written down.

    $29

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned
    Persistent Agent Memory

    Persistent Agent Memory

    Example session with this skill installed

    My agent keeps forgetting decisions we already made, and I have six months of conversations spread across chats. How do I make it remember, and find the related stuff when I ask?

    • Read your context and instructions
    • Compiled the persistent agent memory

    You get a memory store you own: a session note for each finished conversation with decisions, facts, preferences, open questions, and next steps, all dated and sourced, plus a search that answers "Harborline renewal" with the service agreement conversation even though those words never appear in it, because the entry shares the client and the renewal tag. You also get a weekly digest of decisions made, facts learned, preferences changed, and open questions still standing, and in Company Brain mode a shared store your team reads from, with private folders that stay private.

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

    About this skill

    A buyer-owned workflow for the forgetting problem: agents say sorry for the oversight and then forget again, because the session ends and nothing durable is written down. After a session, you capture what was decided, learned, preferred, and left open, with a date and a source on every item, and file it in a plain folder of markdown files you own. You link related memories to each other, and later retrieval scores memories on keyword overlap, shared entities and tags, and recency, so it finds related memories even when the words do not match. It includes a standard-library-only Python search script, memory templates for session notes, decision logs, people files, and preferences, a retrieval rubric with worked examples, and a setup guide for a shared Company Brain mode where a team of 2 to 10 keeps one context, with privacy rules that keep one person's private facts out of another person's retrieval. Built from three real asks: a developer whose agents keep forgetting what was said and decided, a user with six months of conversation archives who wants retrieval that connects conversations that look irrelevant and surfaces patterns, and a CTO who asked for an AI company brain that keeps the context of what is going on so different roles can use agents to draft emails, prep meetings, update client info after meetings, and query how a system is working for a client. Guardrails included: it never invents a memory, never states a memory as fact without its date and source, never silently overwrites a contradiction, and never stores passwords, secrets, payment details, or medical information. Works with Claude Code, Cursor, Codex CLI, GitHub Copilot, and Gemini CLI. The memory store is a folder of markdown files on your own machine. No cloud service, no subscription, no accounts or API keys.

    Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Agents & Orchestration.

    It scores local markdown memory files only; it does not capture transcripts automatically, the user provides the session or its summary. Retrieval finds memories, it does not verify them; a memory is only as true as its source. Company Brain privacy is enforced by folder layout and retrieval rules, not by code access control; a real access control system would need a hosted service. It does not transcribe voice recordings on its own; transcripts and notes go into the store as text.

    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 buying a skill.

    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 5 days ago

    Software
    Python

    Permissions

    Terminal / Shell
    Read Files
    Write Files
    Listed5 days ago

    What's inside

    Frequently Asked Questions

    The 'agent memory' skill provides a durable Markdown-based storage layer that allows agents to record facts and decisions, preventing them from repeating questions or losing context after a session ends.

    This skill is built with Python 3 and uses standard Markdown files, making it compatible with any AI agent or framework capable of executing local scripts and reading/writing text files.

    No, this is a local-first solution that requires no external databases or API keys; all memory is stored in human-readable Markdown files on your own infrastructure.

    You will receive the core Python scoring and retrieval scripts, a set of structured Markdown templates for memory capture, and a setup guide to integrate the system into your agent's workflow.

    The 'Company Brain Mode' allows teams of 2 to 10 to share a unified directory of Markdown files, using a built-in scoring algorithm to ensure agents retrieve the most relevant, up-to-date team context.

    While ideal for facts and decisions, it is not a secure vault; you should not use this skill to store passwords, API secrets, or highly sensitive personally identifiable information (PII).

    New in Agents & Orchestration