rcode

    1

    You re-explain the stack, the decisions, what broke yesterday.

    Free

    Security scanned

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    About this skill

    Every new AI session starts from zero. You re-explain the stack, the decisions, what broke yesterday. rcode ends that.

    It installs a git-tracked Memory Bank (.rcode/memory/) and a team of specialist agents into your IDE. Plain markdown files: no daemon, no vector DB, no runtime dependencies.

    Try it in 60 seconds

    npx @hanzlaa/rcode install
    /rcode-init     # 5 questions, your Memory Bank is born
    /rcode-next     # rcode reads project state and tells you the best next step
    

    Close the session, come back tomorrow, run /rcode-status: it still knows everything.

    What you get

    • Memory that survives sessions: stack, decisions, glossary, known issues and milestones, versioned with your code.
    • A team, not a chatbot: Waleed (CTO), Yousef (backend), Haitham (frontend/RTL), Fatima (QA), Hussain (PM), Sadiq (strategy), plus UX, branding, ML, docs and growth specialists.
    • /rcode-council, the fun one: ask "rewrite or refactor?" and 3-5 specialists answer in parallel. You get a synthesis with the dissent preserved. No more yes-man AI.
    • Plan, execute, verify, ship: /rcode-plan then /rcode-execute (parallel executors, atomic commits) then /rcode-verify-phase, a goal-backward check that the code delivers what the phase promised.
    • Hard-nosed reviewers: /rcode-code-review, /rcode-debug (scientific method, escalates after 3 failed hypotheses), /rcode-harden, /rcode-prove-it.

    Fun first experiments

    1. /rcode-map-codebase, then talk to Waleed (/rcode-waleed-architect) about what is fragile.
    2. Settle a debate: /rcode-council "Should we move to a monorepo?" and read the dissent.
    3. Let it run: /rcode-autonomous plans, executes and verifies remaining phases, pausing only at checkpoints and failures.
    4. Watch it think: /rcode-dashboard opens the Diwan view-only dashboard with phases, decisions and council sessions.

    Works with

    Claude Code (primary), Cursor, Gemini, VS Code, Antigravity.

    Known limitations

    Needs a one-time npx install. Commit .rcode/memory/ to git so memory follows your team across machines.

    Free and MIT licensed: https://github.com/hanzlahabib/rihal-code

    Requires manual installation via npx. Memory Bank files must be committed to git to persist across environments. Primarily optimized for Claude Code.

    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 7 days ago

    • Free with an Agensi account

    Requires

    @hanzlaa/rcode
    Node.js

    Permissions

    Terminal / Shell
    Read Files
    Listed7 days ago

    What's inside

    Frequently Asked Questions

    It solves 'context drift' by creating a git-tracked directory (`.rcode/memory/`) that acts as a permanent brain for your AI, storing architectural decisions, tech stacks, and progress so you never have to re-explain your project in new sessions.

    While optimized for Claude Code, the Memory Bank uses standard Markdown files, making it compatible with Cursor, VS Code, and Gemini-based agents that can read your local workspace.

    The Decision Council allows you to run parallel analysis from different expert personas (like Frontend vs. Architecture), providing a synthesized recommendation that highlights trade-offs instead of just giving a single generic answer.

    You will receive the full suite of specialized agent personas, the automated command set for planning and execution, and the standardized Memory Bank templates required to initialize your workspace.

    Setup is handled via a simple `npx` command that initializes the folder structure and configuration files directly in your repository, ensuring all AI agents recognize the new workflow immediately.

    Yes, because the Memory Bank is stored as Markdown files within your repository, it is version-controlled; you can track how your project's architecture and goals evolve over time just like your code.

    New in Agents & Orchestration

    Free