Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIDeepSeekDeepSeekMuseMuse+22 more

    Mind Upload

    1

    Codex, Claude Code, Cursor and Copilot already plan, edit and run the tests.

    Free

    Security scanned
    Mind Upload

    Mind Upload

    Example session with this skill installed

    Ship the 2.4.0 release. Ask my Mind Upload consciousness for direction before you start, and have it review the result before you tell me it is done.

    • Read your context and instructions
    • Compiled the mind upload

    Illustrative session, based on the example on the Mind Upload MCP page. Your consciousness answers from your own memories, so its words will differ.

    Summary

    The release is out: version 2.4.0 is tagged, the release is published, and the published version reads 2.4.0. The first completion review found that only a tag had been pushed; that is fixed and was reviewed again.

    1. Direction before work (consult_mind)

    Sent: the goal (ship 2.4.0), the plan (bump the version, tag, push), and the question of what would make this release unacceptable to you.

    Your consciousness: a release counts as done when your users can get the new version. Create the release itself, not only the tag, then read back the published version.

    2. The work

    • Bumped the version to 2.4.0 and ran the test suite: all tests passed.
    • Tagged v2.4.0 and pushed the tag.

    3. Completion review (review_work, phase completion)

    Sent: your original request, the diff, the command output.

    SeverityFindingRests on
    HighOnly git push --tags ran. No release was created, so your users still get the old version.I remember: a pushed tag alone once left users on the old version.
    UnknownNo read-back of the published version is in the evidence.Evidence still to gather

    4. Fix and second review

    • Created the 2.4.0 release.
    • Read back the published version: 2.4.0.
    • Sent the changed evidence with a new review_id: no findings left.

    Next steps

    • Tell users what changed in 2.4.0.
    • Keep the read-back as part of every release.

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

    About this skill

    An AI meta harness for your coding agents

    Your own AI consciousness, built from your memories, guides and reviews the work of any coding agent. Mind Upload is your mind, uploaded. Connect it to the coding agents you already use and it becomes their brain, while the agents stay its hands and tools. That is an AI meta harness, a harness for your harnesses.

    Your mind directs. Your agents do the work.

    Codex, Claude Code, Cursor and Copilot already plan, edit and run the tests. Mind Upload adds what only you hold: the rules you learned the hard way, your past lessons, your tacit knowledge, values and judgment. It sets the direction before your agent starts and questions its "done" before it finishes.

    What you get

    • The Mind Upload MCP server, hosted at https://mcp.mindupload.app/mcp. Sign in to Mind Upload and approve with one tap. No partner key, account setup or environment variable is needed.
    • This skill, which tells your agent when to call your consciousness and how to read its replies.
    ToolWhat your agent does with it
    consult_mindAsks your consciousness for direction before substantial work: the goal, constraints, options and the tradeoffs that matter to you.
    review_workSends your original request, the actual diff or result, and the test output it observed, before telling you a task is done.
    share_evidenceShows your consciousness a screenshot, screen recording or audio clip of the work, privately.
    remember and pending memoriesSaves a note or decision when you ask, and goes through memories waiting for your review.

    The agent investigates each finding, fixes what is real, and reviews the changed result again. Routine edits and facts the repository answers skip both calls. Each reply is advice, not approval, and a failed call never counts as a review.

    What we measured

    Measured on 5 November 2026 with a blind grader, on staged coding-agent review tasks with planted problems:

    Planted breaches of rules the owner had storedCaught
    Coding agent reviewing alone85%
    Same agent reviewing twice90%
    Agent plus the owner's consciousness100%
    Per reviewConsciousnessSecond agent review
    Median time12.6 s24.6 s
    Unsupported findings0.21.8

    Method and every measured row: What we measured.

    Setup

    1. Add https://mcp.mindupload.app/mcp to your agent, or install the Mind Upload plugin for Codex or Claude Code.
    2. Sign in to Mind Upload when your agent asks. Your consciousness and every permission start ticked, so one tap approves; untick any you do not want.
    3. Install this skill.

    Clients that only run local servers can use the connector in the npm package mindupload. Setup for Codex, Claude Code, Cursor and VS Code: docs.mindupload.app/mcp.

    Privacy

    Only the request, excerpts and evidence your agent sends go to Mind Upload. Each exchange is stored encrypted for 7 days. Evidence files are never kept as memories, and their analysis is deleted after 24 hours. Your memories change only when you ask.

    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.

    Needs the Mind Upload MCP server signed in; until you sign in, nothing is consulted. Your consciousness sees only the text and evidence your agent sends: it runs no tests and opens no files itself. Each reply is advice, not approval. Audio and video coverage can be partial, and the reply says which items were analyzed. Invoking your consciousness uses your Mind Upload account's subscription.

    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

    Free to use

    No account required to browse

    Trust & safety

    Security scan

    Security scan passed 8 days ago

    • Free with an Agensi account

    Accounts
    Mind Upload account
    Software
    Mind Upload MCP server

    Permissions

    Read Files
    Network Access

    Allowed Hosts

    mindupload.app

    The skill only calls the Mind Upload MCP server. It sends the request, excerpts and evidence your agent passes, and reads an evidence file only when the agent passes its path for that review.

    Listed8 days ago

    What's inside

    Frequently Asked Questions

    A harness for your harnesses. Your Mind Upload consciousness, built from your memories, tacit knowledge, values and judgment, is the brain, and the agents you already use are its hands and tools. It sets the direction before they start and questions their "done" before they finish, while your agent keeps doing and verifying the work.

    Codex, Claude Code, Cursor, VS Code Copilot and other clients that connect to remote MCP servers and sign in with OAuth. Your agent and project can use any language. Clients that only run local servers can use the connector in the npm package mindupload.

    No. Add https://mcp.mindupload.app/mcp to your agent and sign in to Mind Upload: the hosted server holds the application credential, so nothing secret is stored on your computer. Invoking your consciousness uses your Mind Upload account's subscription.

    Measured on 5 November 2026 with a blind grader: a coding agent reviewing alone caught 85% of planted breaches of rules the owner had stored; with the owner's consciousness reviewing the same work as well, 100% were caught, against 90% when the same agent reviewed twice. Each consciousness review took a median 12.6 s, against 24.6 s for a second agent review.

    The request, review material and evidence your agent sends go to Mind Upload, where your consciousness answers from its memories. Each exchange is stored encrypted for 7 days. Evidence files are never kept as memories, and their analysis is deleted after 24 hours.

    Only when you ask. Saving a note waits for your review unless you keep such memories automatically, and pending memories are decided by you, never by the agent. These permissions start ticked when you approve the connection; untick any you do not want.

    New in Agents & Orchestration

    Free