Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    Codex Next Prompt

    1

    Your next prompt, ready to paste: follow-ups for unfinished work, copied to your clipboard.

    Free

    1 useSecurity scanned

    Codex Next Prompt

    Example session with this skill installed

    Write a two-part README outline for a small command-line timer. Part one should cover the introduction and basic use; part two should cover configuration and troubleshooting. Write only part one for now.

    • Read your context and instructions
    • Compiled the codex next prompt

    Illustrative session with the hooks activated

    Below is an example of the workflow; exact wording varies by conversation and model.

    Codex reply

    Timer CLI

    A small command-line timer for focused work sessions, short breaks and repeatable routines. Start a timer from your terminal and keep working until the session ends.

    Part one: introduction and basic use

    Introduction

    Explain what the timer does, who it is for, and what the user needs before running it. Keep the opening focused on the core workflow: choose a duration, start the session, and watch the remaining time.

    Basic use

    Show how to start a timer with the default duration, choose a custom duration, and stop an active session. Include one concise example for each action using the actual commands supported by the project.

    Part two, covering configuration and troubleshooting, is still pending as requested.

    → Want me to “write part two of the README outline, covering configuration and troubleshooting”?

    What goes onto the clipboard

    write part two of the README outline, covering configuration and troubleshooting
    

    Only the quoted instruction is copied. With notifications enabled, the plugin also announces the copy. The user can review the prompt, paste it with Ctrl+V or Cmd+V, and choose whether to send it. Nothing is automatically submitted or executed. If the requested work were already complete, a suggestion would normally be omitted.

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

    About this skill

    Your next prompt, ready to paste

    Codex Next Prompt is a free, MIT-licensed plugin for the Codex app and Codex CLI. When work you already requested is unfinished, Codex can end its reply with one natural sentence quoting a useful follow-up. The plugin copies only that quoted instruction to your clipboard, ready for Ctrl+V or Cmd+V. You review, paste and send it yourself.

    What it does

    • Suggests follow-ups for unfinished work, an outstanding part of your request, or returning to work after a side question.
    • Skips suggestions when a task is complete or you ask to stop.
    • Rotates phrasing and follows your conversation language.
    • Copies the quoted prompt and optionally shows a desktop notification. Both settings can be changed.
    • Includes setup, status, diagnosis, enable/disable and manual suggestion commands.

    Install and activate

    The download includes the complete plugin source, Python runtime code, hook definitions, platform installers and eight SKILL.md entrypoints. Run the included Windows PowerShell or macOS/Linux installer, then use $nextprompt-onboarding. Fully quit and reopen Codex and review the SessionStart, UserPromptSubmit and Stop hooks in Codex Hook settings (CLI: /hooks). See README.md for the full instructions.

    Compatibility

    Automatic suggestions and clipboard copy require OpenAI Codex with plugin hooks (Codex CLI 0.159+ or a compatible Codex app), Python 3.9+ and a working clipboard bridge. Windows, macOS and common Linux desktops are supported; WSL clipboard behavior depends on the host bridge. A generic SKILL.md agent does not gain automatic Codex hook support simply by downloading the package.

    Privacy and control

    Suggestions are never automatically sent or executed. Recognizable credentials are redacted, and lines that look like secrets are not copied. Local usage statistics contain counts and short hashes rather than conversation text, and can be disabled. The default inline mode uses your active Codex model without a separate model request; optional model mode uses your configured provider.

    Free and open source

    Version 0.2.4, distributed under the MIT license. This skill has no purchase or subscription charge. Your own Codex access and any model usage are subject to your service plan.

    Source: https://github.com/Jineggegg/codex-next-prompt

    The gallery image and the sample below illustrate the workflow; exact wording varies by conversation and model.

    Automatic operation requires Codex plugin hooks; downloading SKILL.md alone is not enough. Hooks need review and trust in Codex after restarting. Clipboard support depends on the OS/WSL bridge. Suggestions are model-generated and may occasionally be omitted or imperfect. Default inline mode has no extra model request, but it uses the active Codex model; optional model mode consumes provider quota. No paid Codex plan or model credits are included.

    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

    1 use

    Times used to date

    Free forever

    No account required to browse

    Trust & safety

    Security scan

    Security scan passed 26 days ago

    • Free with an Agensi account

    Requires

    Codex app or Codex CLI 0.159+
    Python 3.9+
    PowerShell or POSIX shell
    Git
    System clipboard bridge
    Python
    macOS

    Permissions

    Terminal / Shell
    Read Files
    Write Files
    Network Access
    Environment Variables

    Shell and file access install and run the bundled Python plugin, read hook context, and save local settings/statistics in Codex plugin data. Environment variables locate plugin/data directories. Clipboard writes and optional desktop notifications occur locally. Network access is used for installation prerequisites and the optional model provider; default inline mode makes no separate provider request. Suggestions are never automatically sent or executed.

    Listed26 days ago

    What's inside

    Frequently Asked Questions

    It suggests a next instruction when a part of your existing request is unfinished, this turn leaves work to do, or you return to unfinished work after a side question. Finished tasks, wrap-up requests and stop requests normally receive no suggestion.

    Automatic suggestions and clipboard copy require the OpenAI Codex app or Codex CLI with plugin hooks, plus Python 3.9+. Windows, macOS and common Linux desktops are supported. Other SKILL.md agents can read the setup entrypoint, but the automatic lifecycle is specific to Codex.

    Yes. This listing is free and the code is MIT-licensed. You receive the complete plugin source, Python runtime code, hook definitions, eight skill entrypoints, installers and documentation. Your own Codex service access and any model usage are separate and follow your account plan.

    Use the platform installer in the package, then run $nextprompt-onboarding. Fully quit and reopen Codex, and review and trust the SessionStart, UserPromptSubmit and Stop hooks in its Hook settings (CLI: /hooks). A successful installation or doctor self-test alone does not prove that hooks are trusted in your current chat.

    Yes. Use $nextprompt-setup to turn automatic copy and desktop notifications on or off. You always decide whether to paste and send a suggestion; the plugin never automatically submits or executes it.

    WSL needs an available clipboard bridge to the host or Linux display. If neither is available, suggestions can be displayed without copying. Windows and WSL may have different Python installations and settings; run hooks on the host where you want clipboard and notifications to work.

    New in Prompt & Skill Engineering