{
  "markdown": "# MAGI — Multi-Perspective Analysis Extension for Gemini CLI\n\n[![Verify](https://github.com/BolivarTech/MAGI-Gem/actions/workflows/verify.yml/badge.svg)](https://github.com/BolivarTech/MAGI-Gem/actions/workflows/verify.yml)\n[![Python 3.9+](https://img.shields.io/badge/python-3.9%2B-blue.svg)](https://www.python.org/downloads/)\n[![Tests](https://img.shields.io/badge/tests-16%20passing-brightgreen.svg)](#tests)\n[![Ruff](https://img.shields.io/badge/linter-ruff-orange.svg)](https://docs.astral.sh/ruff/)\n[![License](https://img.shields.io/badge/license-MIT%20OR%20Apache--2.0-blue.svg)](#license)\n\nA Gemini CLI extension that implements a **multi-perspective analysis system** inspired by the [MAGI supercomputers](https://evangelion.fandom.com/wiki/Magi) from *Neon Genesis Evangelion*.\n\nThree specialized AI agents independently analyze the same problem from complementary — and deliberately adversarial — perspectives, then synthesize their verdicts via weight-based majority vote.\n\n---\n\n## Why Three Adversarial Perspectives?\n\n### The MAGI in Evangelion\n\nIn *Neon Genesis Evangelion* (1995, Hideaki Anno / Gainax), the MAGI are three supercomputers that govern Tokyo-3's critical decisions. Each embodies a different facet of their creator, Dr. Naoko Akagi: **Melchior** (the scientist), **Balthasar** (the mother), and **Caspar** (the woman). Decisions require consensus — no single perspective dominates.\n\n### The Theory in Practice\n\nThe adversarial multi-perspective model addresses well-documented cognitive biases in software engineering:\n\n| Bias | How MAGI Mitigates It |\n|------|----------------------|\n| **Confirmation bias** | Three agents with different evaluation criteria are unlikely to share the same blind spots |\n| **Anchoring** | Agents analyze independently — no agent sees the others' output before forming its own verdict |\n| **Groupthink** | Caspar (Critic) is designed to be adversarial; its role is to find fault, not agree |\n| **Optimism bias** | The weight-based scoring penalizes reject (-1) more heavily than approve (+1), making negative signals harder to override |\n\nThe key insight is that **disagreement between agents is a feature, not a failure**. When Melchior (Scientist) approves but Caspar (Critic) rejects, the dissent surfaces a genuine tension between technical correctness and risk tolerance.\n\n---\n\n## Agents\n\n| Agent | Codename | Lens | Personality |\n|-------|----------|------|-------------|\n| **Melchior** | Scientist | Technical rigor and correctness | Precise, evidence-based, favors proven solutions |\n| **Balthasar** | Pragmatist | Practicality and maintainability | Grounded, trade-off oriented, advocates for the team |\n| **Caspar** | Critic | Risk, edge cases, and failure modes | Adversarial by design, finds what others miss |\n\n---\n\n## Installation\n\n### Official Method (Recommended)\n\nInstall the extension directly from the official repository with auto-updates enabled:\n\n```powershell\ngemini extensions install https://github.com/BolivarTech/MAGI-Gem.git --consent --auto-update\n```\n\n### Development Method\n\nIf you want to contribute or modify the extension:\n\n```powershell\ngit clone https://github.com/BolivarTech/MAGI-Gem.git\ncd MAGI-Gem\ngemini extensions install . --consent\n```\n\n---\n\n## Updating\n\nTo update the extension to the latest version (e.g., v0.1.9) from the official repository:\n\n```powershell\ngemini extensions update magi\n```\n\nIf you are using the **Development Method**, just pull the latest changes and reinstall:\n```powershell\ngit pull origin main\ngemini extensions uninstall magi\ngemini extensions install . --consent\n```\n\n---\n\n## Usage\n\nInvoke with `gemini magi` or trigger the skill naturally:\n\n```bash\ngemini magi \"review this code\"\ngemini \"Give me three perspectives on this design\"\n```\n\n### Modes\n\n| Mode | When to Use | Example |\n|------|-------------|---------|\n| `code-review` | Reviewing code or diffs | \"magi review this PR\" |\n| `design` | Evaluating architecture decisions | \"magi analyze this migration plan\" |\n| `analysis` | General problem analysis, trade-offs | \"magi should we use Redis or Postgres?\" |\n\n---\n\n## How It Works\n\n1. **Progressive Disclosure** — The `SKILL.md` identifies the need for deep analysis.\n2. **Parallel Dispatch** — The orchestrator `run_magi.py` launches 3x `gemini -p` calls in parallel using your existing CLI authentication.\n3. **Synthesis** — Agent outputs are validated, weighted, and synthesized into a final report.\n\n---\n\n## Project Structure\n\n```\ngemini-extension.json         -- Extension manifest\nskills/magi/\n  SKILL.md                    -- Skill entry point and workflow\n  agents/                     -- Agent system prompts (.md)\n  scripts/                    -- Core Python logic (consensus, reporting, etc.)\n```\n\n---\n\n## License\n\nThis project is licensed under the MIT License - see the [LICENSE](LICENSE) file for details.\n\n---\n\n## Credits\n\nThe MAGI concept originates from *Neon Genesis Evangelion* (1995) by Hideaki Anno / Gainax. This extension is a creative adaptation of that decision-making philosophy for modern software engineering.\n",
  "bytes": 5078,
  "sha": "d38d30307fe5991feb9d11f1e3c2658bd42691b42cc02f99f13a928a6bec97c9",
  "repo_slug": "bolivartech/magi-gem",
  "fonte": "repo",
  "truncated": false,
  "api": "https://agentalog.com/api/listings/plg_bolivartech_magi_gem_1b29b9de/readme"
}