More screenshots

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    DSA Coach

    2

    Turns your agent into a DSA coach on any coding platform: plans practice around your weakest topics and repeated mistakes, keeps a mistake journal, and schedules revisions with FSRS spaced repetition.

    Free

    Security scanned
    DSA Coach

    DSA Coach

    Example session with this skill installed

    Plan my DSA practice for today. My LeetCode username is axryaa.

    • Read your context and instructions
    • Compiled the dsa coach

    Tool data below is from a real run against LeetCode on 3 Nov 2026; the agent's wording varies by model.

    Where you stand

    I read your LeetCode profile with get_profile:

    LevelSolved
    Easy140
    Medium99
    Hard8
    • Strongest tags: array (133), hash-table (46), two-pointers (42)
    • Thinnest tags: game-theory, randomized, design and queue, with 1 problem each

    Your mistake journal is empty, so today's plan comes from your tag gaps alone. Once you log a few reviews, plans will also target the mistakes you repeat.

    Today's practice

    I searched with search_problems for Easy game-theory problems you haven't solved yet:

    ProblemLevelWhy this one
    Nim GameEasyGame theory is your thinnest tag; Nim is the classic way in
    Divisor GameEasySame idea, winning and losing positions, in a new setting

    Start at Easy: with one game-theory problem solved, the pattern matters more than difficulty. When you can explain why the first player wins, move up to Medium.

    Next steps

    1. Solve Nim Game, then tell me how it went: again, hard, good or easy, plus anything you got wrong.
    2. I'll log it with log_review, record the mistakes in your journal, and FSRS will schedule your first revision.
    3. Ask for revisions_due or a quiz any time to see what to revise.

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

    About this skill

    DSA Coach pairs this skill with the open-source dsa-coach MCP server. The server provides the data; the skill tells your agent when to use each tool and how.

    What it does

    • Any platform: works with whatever judge you practise on. Platforms with an adapter (LeetCode and Codeforces today) are also read from their APIs: your solved-by-tag counts and levels, and problems you haven't solved yet.
    • Practice plans built from your weak topics and repeated mistakes. Each problem names the weakness it targets.
    • Mistake journal: every retry records what went wrong, in your own words, and the agent reuses the same labels so patterns show up.
    • FSRS spaced repetition: revisions come due when you are about to forget a problem. After a few hundred reviews, the schedule fits to your own memory.
    • Platforms without an adapter: a problem is recorded once from its page, through a browser MCP server, then reviewed and revised like any other.
    • Dashboard: progress, due revisions and a mistake heatmap, rendered in the chat by hosts that support MCP Apps.

    Setup

    Install uv, then add the server:

    claude mcp add dsa-coach -- uvx dsa-coach-mcp
    

    In Cursor and other MCP clients, set the command to uvx and the args to ["dsa-coach-mcp"].

    Network and data

    The server calls LeetCode's public GraphQL API and Codeforces' official REST API to read public profiles and problems. Your reviews and mistakes stay in a local SQLite file. No account, API key or telemetry.

    Open source, MIT license: https://github.com/aaryagodbole/dsa-coach-mcp

    Needs the dsa-coach MCP server and uv. LeetCode and Codeforces have adapters; other platforms are recorded from their pages and cannot be searched for new problems. Fitting the schedule to your own memory needs a few hundred reviews.

    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

    Free forever

    No account required to browse

    Trust & safety

    Security scan

    Security scan passed 14 days ago

    • Free with an Agensi account

    Requires

    Dsa Coach MCP
    uv (Python tool runner)
    Playwright (Optional) MCP

    Permissions

    Network Access
    Listed14 days ago

    What's inside

    Frequently Asked Questions

    Every review is rated again, hard, good or easy and logged. FSRS estimates how likely you are to still recall each problem, and the revision list puts the one you are most likely to have forgotten first.

    Yes, from public data: given your handle, it reads how many problems you solved per tag and level, and it can exclude problems you already solved when searching. Revision scheduling starts from the reviews you log with the coach.

    uv, and the dsa-coach MCP server: claude mcp add dsa-coach -- uvx dsa-coach-mcp. No account or API key is needed.

    Yes. With a browser MCP server connected, the agent reads the problem page and records it once; after that it is reviewed and revised like any other problem. Such platforms cannot be searched for new problems.

    A plain prompt forgets everything after the chat. This keeps your reviews and mistakes in a local database, picks problems you have not solved from your real profile, and tells you when to revise each one.

    New in Education & Learning