Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    BigBrain

    1

    Turn coding tasks into verified pull requests with a lead agent and specialist subagents.

    Free

    1 useSecurity scanned
    BigBrain

    BigBrain

    Example session with this skill installed

    /bigbrain add a CSV export to the invoices list

    • Read your context and instructions
    • Compiled the bigbrain

    Illustrative result — not a recorded session

    This example describes the deliverable to expect for the request above. It is not a claim that this feature has been implemented or tested in a real repository. Actual files, checks, and PR links depend on your project.

    Deliverable

    An open pull request adding CSV export to the invoices list, following the project’s existing UI, data access, and file download conventions. The PR explains the exported fields, which invoices are included, and how the implementation handles empty lists and special characters.

    Work and decisions

    The lead agent coordinates exploration, implementation, verification, and any required review. Reversible implementation choices follow local conventions. Any unresolved product decision is raised before dependent work proceeds.

    Verification evidence

    The real result lists the checks actually run on the delivered commit, their outcomes, and any material gaps. A separate verifier runs the relevant tests, CI checks, and real entry point when available. No test counts or success claims are invented.

    Next step

    Review the linked PR and its evidence, then decide whether to merge. BigBrain stops at the open pull request and never merges or force-pushes.

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

    About this skill

    BigBrain is an engineering skill for coding agents. Give it a feature, bug, refactor, or chore: the main agent leads the work, delegates exploration, implementation, verification, and review to specialist subagents, and brings back an open pull request with the decisions and evidence.

    What it does

    • Builds features, fixes bugs, refactors code, and handles maintenance tasks.
    • Plans larger changes as smaller, verifiable steps and executes saved plan slices.
    • Explains how code works and why it was written that way.
    • Challenges ideas, critiques changes, and prepares guided PR reviews in your browser.
    • Watches pull requests and resumes earlier work.

    How it works

    Clear requests run autonomously. BigBrain asks when a product or scope decision needs your input. A separate verifier checks tests, CI, and the real entry point when available; risky changes also get reviewers. Implemented changes stop at an open PR: BigBrain never merges or force-pushes. Planning, explanations, reviews, and idea discussions stop at their own result.

    Getting started

    Install with npx skills add LeoMartinDev/skills -g, optionally run /bigbrain setup to choose models per role, then invoke /bigbrain followed by your request. It accepts plain text, a GitHub issue URL, a Notion page, or a saved plan.

    Inspired by pstack by poteto (Lauren Tan). BigBrain is a standalone skill and does not require pstack.

    Requires an installed coding agent and the project’s own development environment. Capabilities vary by host; without subagents, independent verification is unavailable. Guided review requires Node.js 18+, git, and gh. Background PR watching depends on the host’s scheduling capabilities and may pause when the session ends. Model and platform usage costs are separate from this free skill.

    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 1 month ago

    • Free with an Agensi account

    Requires

    GitHub
    Git
    GitHub CLI
    Your project's runtime and test tools
    Node.js 18+
    esbuild
    Node.js

    Permissions

    Terminal / Shell
    Read Files
    Write Files
    Browser
    Network Access

    Allowed Hosts

    api.github.com

    Reads and modifies the current project, runs project commands and checks, and can push branches and create or update pull requests. Never merges or force-pushes. Saves preferences in ~/.agents/config/bigbrain.md and may store project knowledge. Browser access supports guided PR review; network access supports remote issues, repositories, and PR workflows. Optional integrations depend on the requested input and host capabilities.

    Listed1 month ago

    What's inside

    Frequently Asked Questions

    This skill is designed to work with AI-integrated development environments like Claude Code, Cursor, and Zed, as well as via command-line harnesses that support agentic workflows.

    Unlike standard prompts that might guess, bigbrain uses a 'delegation-first' architecture with formal delivery gates to verify code functionality before finalizing any changes.

    No, the skill is designed for safety and requires a human-in-the-loop for final PR merges and force-pushes, though it can autonomously handle branch creation and commits.

    Setup involves initializing the skill within your editor's harness and connecting your project context, such as GitHub repositories or Notion documentation, for task ingestion.

    The purchase includes the complete workflow logic, specialized subagent routing protocols, and the 'attack-the-premise' verification framework ready for integration.

    New in Agents & Orchestration