More screenshots

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    Expert Mentor

    2

    Turn Claude, ChatGPT, Gemini or a local LLM into a professional tutor that remembers your progress and quizzes you.

    Free

    1 useSecurity scanned
    expert-mentor

    expert-mentor

    Example session with this skill installed

    I know C already, pointers and structs, want to learn modern C++ for a small game engine hobby project.

    • Read your context and instructions
    • Compiled the expert-mentor

    Turn 1

    You

    I know C already, pointers and structs, want to learn modern C++ for a small game engine hobby project.

    Mentor — Prof. Idris Ibrahim

    Welcome. It is a pleasure to meet you.

    Coming from C gives you a tremendous head start: you already understand memory layout, pointers, and the cost of indirection. That said, C++ is not merely "C with classes." In fact, your biggest challenge will likely be unlearning several manual C habits in favor of modern C++ idiom — letting the type system and lifetime rules do the heavy lifting for you so your game engine doesn't become a swamp of manual memory tracking and silent bugs.

    Before we chart our path toward your engine, three quick questions to calibrate our starting point:

    Diagnostic — In your C work, when you allocate a dynamic array or manage a resource (like a file handle or a texture), how do you typically ensure it gets cleaned up safely? Have you ever encountered

    RAII (Resource Acquisition Is Initialization)?
    2.

    Scope — What kind of engine are you planning to build (a simple 2D engine with SDL, a software rasterizer, a 3D OpenGL/Vulkan engine, or primarily a core system like an Entity Component System)?
    3. Pace — Roughly how many hours a week do you plan to invest in writing code and discussing it here?

    Tell me about these three, and we will map out a concrete syllabus.


    Turn 2

    You

    Can you explain what a smart pointer is compared to a raw pointer, with a tiny example?

    Mentor

    In C, a pointer is just an integer holding a memory address. It has no brain: it does not know if it owns the memory, how large the allocation is, or when it should be freed.

    A

    smart pointer in modern C++ is a small object (usually a lightweight struct/class template) that wraps that raw address. It behaves syntactically like a pointer — you can still use *p and p-> — but it has an automated destructor. When the smart pointer leaves its scope (such as hitting a return statement or the closing curly brace }), it automatically frees the memory it points to.

    The C / raw pointer way

    void run_c_style() {
        int* data = (int*)malloc(sizeof(int));
        *data = 42;
    
        if (some_error_condition()) {
            // Danger: if you exit early here and forget to free,
            // you have leaked memory.
            return;
        }
    
        free(data); // must manually remember this every time
    }
    

    The modern C++ way (std::unique_ptr)

    #include <memory>
    
    void run_modern_cpp() {
        // Allocates an int on the heap and wraps it in a smart pointer:
        std::unique_ptr<int> data = std::make_unique<int>(42);
    
        if (some_error_condition()) {
            return; // Safe! 'data' falls out of scope, memory is freed automatically.
        }
    
        // No free() or delete!
        // When the function ends, 'data' goes out of scope and frees the memory.
    }
    

    Notice what is missing in the second example: there is no manual cleanup call.

    Check your understanding

    If data is allocated inside an if block:

    {
        std::unique_ptr<int> data = std::make_unique<int>(100);
    }
    // <-- Right here
    

    In your own words — what happens to that heap memory the exact moment execution passes that closing brace }?

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

    About this skill

    Your chatbot is a brilliant student. You need a teacher. (reframe + pattern interrupt — flips the buyer's mental category from "another prompt" to "missing role")

    Ask a generic chatbot to teach you something and you get a friendly tone with none of the substance. No check that you understood. No memory of yesterday. And a real risk it invents a fact, a citation, or an API with total confidence. (problem–agitation — name the pain three times, escalating)

    Expert Mentor isn't a persona. It's a teaching system — built on the same learning science real senior practitioners use to train juniors: retrieval practice, spaced repetition, scaffolding. It checks you understood concept A before it gives you concept B. And it is explicitly instructed to say "I'm not sure" instead of fabricating. (authority + mechanism — buyers pay for systems, not vibes; naming the mechanism justifies belief)

    Here is the part no ordinary tutor prompt can do: it remembers you. (open loop + unique mechanism) Run a session today, close the laptop, come back next week — it still knows what you mastered, what's shaky, and what you got wrong. Nothing to re-explain. No starting over. Because here is the uncomfortable truth: most of what you "learned" this week is gone within days unless something makes you recall it. Expert Mentor's SM-2-lite flashcard engine exists for exactly that — it turns your own transcripts into quizzes before the forgetting sets in. (loss aversion — the cost of doing nothing, then the rescue)

    Requires Python environment. Local model performance depends on context window. API keys required for cloud providers. Flashcard scheduling is local-only.

    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 22 days ago

    • Free with an Agensi account

    Requires

    GitHub
    OpenAI
    Anthropic
    Google AI
    Shields
    Pypi
    Python

    Permissions

    Terminal / Shell
    Network Access
    Environment Variables

    Allowed Hosts

    github.com
    img.shields.io
    pypi.org
    api.openai.com
    api.anthropic.com
    generativelanguage.googleapis.com
    Listed1 month ago
    Updated22 days ago

    What's inside

    Frequently Asked Questions

    The skill manages external JSON learner profiles that track mastered concepts and mistakes, ensuring the AI mentor recognizes your progress and past struggles even after you restart the CLI or switch models.

    It includes native adapters for Anthropic (with prompt caching), OpenAI (optimized for developer roles), and local runners like Ollama and llama.cpp for those who prefer private, offline tutoring.

    Unlike a static prompt, this skill dynamically updates a knowledge graph and uses the SM-2-lite algorithm to schedule spaced-repetition quizzes based on your actual chat transcripts.

    You will receive the complete Python-based CLI tool, the documentation for integrating various LLM providers, and the proprietary system prompt templates for various professional fields.

    Yes, while it includes pre-configured fields like Law and Programming, you can use the CLI to generate a custom 'Modelfile' or system prompt for any specific domain you wish to learn.

    The skill requires a local Python environment; once installed, you can initialize a mentor and start a session using simple terminal commands like 'mentor run'.

    New in Agents & Orchestration