{"id":51951724,"url":"https://github.com/dubsopenhub/havoc-hackathon","last_synced_at":"2026-07-29T06:01:29.144Z","repository":{"id":339493025,"uuid":"1162164516","full_name":"DUBSOpenHub/havoc-hackathon","owner":"DUBSOpenHub","description":"Pit AI models against each other. Score them sealed. Crown a winner. All built using the GitHub Copilot CLI. ⚡ ","archived":false,"fork":false,"pushed_at":"2026-06-18T18:23:32.000Z","size":18338,"stargazers_count":5,"open_issues_count":1,"forks_count":3,"subscribers_count":0,"default_branch":"main","last_synced_at":"2026-07-07T22:21:29.414Z","etag":null,"topics":["adversarial-ai","ai-agents","blind-adjudication","copilot-cli","copilot-extensions","multi-agent","multi-model","orchestration","prompt-engineering"],"latest_commit_sha":null,"homepage":"","language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"mit","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/DUBSOpenHub.png","metadata":{"files":{"readme":"README.md","changelog":"CHANGELOG.md","contributing":"CONTRIBUTING.md","funding":null,"license":"LICENSE","code_of_conduct":"CODE_OF_CONDUCT.md","threat_model":null,"audit":null,"citation":null,"codeowners":".github/CODEOWNERS","security":"SECURITY.md","support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null,"zenodo":null,"notice":null,"maintainers":null,"copyright":null,"agents":"AGENTS.md","dco":null,"cla":null}},"created_at":"2026-02-19T23:57:16.000Z","updated_at":"2026-05-21T19:27:51.000Z","dependencies_parsed_at":null,"dependency_job_id":null,"html_url":"https://github.com/DUBSOpenHub/havoc-hackathon","commit_stats":null,"previous_names":["dubsopenhub/havoc-hackathon"],"tags_count":1,"template":false,"template_full_name":null,"purl":"pkg:github/DUBSOpenHub/havoc-hackathon","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/DUBSOpenHub%2Fhavoc-hackathon","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/DUBSOpenHub%2Fhavoc-hackathon/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/DUBSOpenHub%2Fhavoc-hackathon/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/DUBSOpenHub%2Fhavoc-hackathon/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/DUBSOpenHub","download_url":"https://codeload.github.com/DUBSOpenHub/havoc-hackathon/tar.gz/refs/heads/main","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/DUBSOpenHub%2Fhavoc-hackathon/sbom","scorecard":null,"host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":286080680,"owners_count":36020486,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2026-07-20T02:08:10.276Z","status":"online","status_checked_at":"2026-07-29T02:00:04.910Z","response_time":95,"last_error":null,"robots_txt_status":"success","robots_txt_updated_at":"2025-07-24T06:49:26.215Z","robots_txt_url":"https://github.com/robots.txt","online":true,"can_crawl_api":true,"host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["adversarial-ai","ai-agents","blind-adjudication","copilot-cli","copilot-extensions","multi-agent","multi-model","orchestration","prompt-engineering"],"created_at":"2026-07-29T06:01:27.390Z","updated_at":"2026-07-29T06:01:29.120Z","avatar_url":"https://github.com/DUBSOpenHub.png","language":"Python","funding_links":[],"categories":[],"sub_categories":[],"readme":"# 🏟️ Havoc Hackathon\n\n**Pit AI models against each other. Score them sealed. Crown a winner. All built using the GitHub Copilot CLI. ⚡**\n\n[![GitHub](https://img.shields.io/badge/GitHub-Copilot_CLI-blue?logo=github)](https://github.com/features/copilot)\n[![License: MIT](https://img.shields.io/badge/License-MIT-green.svg)](LICENSE)\n[![Security Policy](https://img.shields.io/badge/Security-Policy-brightgreen?logo=github)](SECURITY.md)\n\n\u003e ### ⚡ One Command. That's It.\n\u003e\n\u003e **Never used the CLI before? No problem.** Follow these 3 steps:\n\u003e\n\u003e **1. Open your terminal**\n\u003e - 🍎 **Mac:** Press `⌘ + Space`, type **Terminal**, hit Enter\n\u003e - 🪟 **Windows:** Press `Win + X`, choose **Terminal** or **PowerShell**\n\u003e - 🐧 **Linux:** Press `Ctrl + Alt + T`\n\u003e\n\u003e **2. Paste this line and press Enter:**\n\u003e ```bash\n\u003e curl -fsSL https://raw.githubusercontent.com/DUBSOpenHub/havoc-hackathon/main/quickstart.sh | bash\n\u003e ```\n\u003e *Already have the CLI? No worries — this detects it and skips straight to adding the skill.*\n\u003e\n\u003e **3. When Copilot opens, type:** `run hackathon`\n\u003e\n\u003e That's it — let the models compete! 🏟️\n\u003e\n\u003e *Requires an active [Copilot subscription](https://github.com/features/copilot/plans). [Get one here →](https://github.com/features/copilot/plans)*\n\n---\n\n\u003cp align=\"center\"\u003e\n  \u003cimg src=\"docs/demo/demo.gif\" alt=\"Havoc Hackathon Demo — terminal preview showing ELO leaderboard, tournament heats, ensemble synthesis, and winner reveal\" width=\"700\" /\u003e\n  \u003cbr /\u003e\n  \u003csub\u003e▲ Preview of a 13-model tournament: ELO leaderboard → 4 heats → finals → ensemble synthesis → winner reveal · \u003ccode\u003epython3 docs/demo/demo.py\u003c/code\u003e to run locally\u003c/sub\u003e\n\u003c/p\u003e\n\n---\n\n## 🤔 What Is This?\n\n**Havoc Hackathon** is a multi-model orchestration skill for the [Copilot CLI](https://docs.github.com/copilot/concepts/agents/about-copilot-cli) that turns your terminal into a competitive arena. Give it any task, code, copy, design, architecture, naming, anything, and it dispatches up to **13 current full-size AI models** in tournament elimination heats, scores them with sealed judge panels, evolves the best ideas between rounds, synthesizes the final output from collective intelligence across all finalists, and tracks persistent ELO ratings so it gets smarter over time. No servers, no API keys, just a skill file and your terminal.\n\n### 💬 The Problem\n\nYou ask one AI model and hope for the best. But one model gives you one perspective with unknown quality  -  no baseline, no comparison, no way to know if the output is great or just okay. Havoc Hackathon lets you throw any idea into the arena and see how multiple AIs attack it independently, then find out which approach actually holds up under sealed evaluation.\n\n### ⚡ What Makes It Different\n\n- 🏟️ **Tournament mode** – up to 13 current full-size models compete in elimination heats, not just 3\n- 🧬 **Evolution between rounds** – Round 2 finalists learn what won Round 1\n- 🗳️ **Ensemble synthesis** – voting merge across all finalists (CONSENSUS/MAJORITY/UNIQUE)\n- 🔒 **Sealed judging** – judges never see which model wrote what\n- ⚖️ **Per-heat judge panels** – 3 judges × N heats running in parallel\n- 🔄 **Self-correcting scores** – rubrics adapt when judges disagree\n- 📈 **Persistent leaderboard** – tracks which models actually deliver over time\n- 🏆 **Full ceremony** – podiums, drumrolls, and dramatic reveals in your terminal\n- 📄 **Zero infrastructure** – no servers, no API keys, no build step\n\n---\n\n## 📦 Install (30 seconds)\n\n### Prerequisites\n\n- [GitHub Copilot CLI](https://github.com/github/copilot-cli) installed\n- An active [Copilot subscription](https://github.com/features/copilot/plans)\n\n### Add the Skill\n\n#### Instant Install (no clone needed) ⚡\n\n```bash\nmkdir -p ~/.copilot/skills/havoc-hackathon ~/.copilot/agents \u0026\u0026 \\\n  curl -sL https://raw.githubusercontent.com/DUBSOpenHub/havoc-hackathon/main/skills/havoc-hackathon/SKILL.md \\\n    -o ~/.copilot/skills/havoc-hackathon/SKILL.md \u0026\u0026 \\\n  curl -sL https://raw.githubusercontent.com/DUBSOpenHub/havoc-hackathon/main/agents/havoc-hackathon.agent.md \\\n    -o ~/.copilot/agents/havoc-hackathon.agent.md \u0026\u0026 \\\n  curl -sL https://raw.githubusercontent.com/DUBSOpenHub/havoc-hackathon/main/agents/grid-medic.agent.md \\\n    -o ~/.copilot/agents/grid-medic.agent.md \u0026\u0026 \\\n  echo \"✅ Installed! Run /skills reload in Copilot CLI, then say: run hackathon\"\n```\n\n#### Full Install (clone the repo)\n\n```bash\ngit clone https://github.com/DUBSOpenHub/havoc-hackathon.git \u0026\u0026 \\\n  mkdir -p ~/.copilot/skills ~/.copilot/agents \u0026\u0026 \\\n  cp -r havoc-hackathon/skills/havoc-hackathon ~/.copilot/skills/ \u0026\u0026 \\\n  cp havoc-hackathon/agents/*.agent.md ~/.copilot/agents/ \u0026\u0026 \\\n  echo \"✅ Havoc Hackathon installed! Run /skills reload in Copilot CLI.\"\n```\n\n\u003cdetails\u003e\n\u003csummary\u003eOther options\u003c/summary\u003e\n\n**Auto-discovery (run from repo):**\n\n```bash\ngit clone https://github.com/DUBSOpenHub/havoc-hackathon.git\ncd havoc-hackathon\ncopilot\n```\n\nThe skill at `.github/skills/havoc-hackathon/` is automatically discovered  -  no copy needed.\n\n\u003c/details\u003e\n\n---\n\n## 🎮 Run Your First Hackathon\n\n### Run a Hackathon\n\n```\n\u003e run hackathon  -  build a REST API for a todo app\n```\n\n```\n\u003e run hackathon  -  review this PR for security issues\n```\n\n```\n\u003e run hackathon  -  design a landing page hero section\n```\n\n### What Happens\n\n1. **🎬 Opening Ceremony**  -  Arena banner, mode selection (Tournament or Classic), model tier (⚡ Standard or 👑 Premium), contestant lineup, scoring rubric\n2. **🏁 Round 1  -  Heats**  -  All models dispatched in parallel across elimination heats with live commentary\n3. **⚖️ Per-Heat Judging**  -  3-judge sealed panels score each heat independently (up to 12 judge agents in parallel)\n4. **🧬 Evolution Brief**  -  Orchestrator summarizes what strategies won each heat (zero extra LLM calls)\n5. **🏁 Round 2  -  Finals**  -  Heat winners compete with Evolution Brief prepended to their prompt\n6. **🏆 Winner Reveal**  -  Drumroll → fireworks → ASCII podium → detailed scoreboard\n7. **🗳️ Ensemble Synthesis**  -  Voting merge across ALL finalists (CONSENSUS/MAJORITY/UNIQUE)\n8. **📈 ELO Update**  -  Per-round leaderboard changes with commentary\n9. **🫡 Closing Ceremony**  -  Tournament bracket recap, final stats, optional replay export\n\n### Customize\n\n- Choose model tier: `\"run hackathon with premium models\"` or `\"run hackathon with standard models\"`\n- Choose specific models: `\"hackathon with opus, gemini, and codex\"`\n- Classic mode (3 models, no heats): `\"run hackathon quick\"` or `\"run hackathon fast\"`\n- Set custom rubric: `\"judge on security, performance, and readability\"`\n- Preflight check: `\"dry run\"` or `\"preflight\"` to validate setup without running a real hackathon\n- Show stats: `\"show leaderboard\"` or `\"show stats\"` anytime\n\n\u003cdetails\u003e\n\u003csummary\u003e\u003cstrong\u003e📼 Example Run  -  Full Hackathon Transcript (Premium Tier)\u003c/strong\u003e\u003c/summary\u003e\n\n```\n\u003e run hackathon  -  write a Mass Effect themed motivational quote for open source developers\n\n╔══════════════════════════════════════════════════════════════════╗\n║              ⚡  H A V O C   H A C K A T H O N  ⚡              ║\n║                                                                  ║\n║  🏟️  THE ARENA IS READY. THE AI MODELS ARE READY TO COMPETE.  🏟️  ║\n╚══════════════════════════════════════════════════════════════════╝\n\n🎯 TASK: Write a Mass Effect themed motivational quote for open source developers\n\n👑 CONTESTANTS:\n  🔵 GPT-5.5               (Premium)\n  🟣 Claude Opus 4.7       (Premium)\n  🔵 Claude Opus 4.6       (Premium)\n\n⚖️ RUBRIC: Clarity · Simplicity · Relevance · Inspiration · Memorability (each /10)\n\n3... 2... 1... GO! 🏁\n\n🏁 ALL MODELS CROSSED THE FINISH LINE!\n  🔵 GPT-5.5             ✅ 3s ⚡ Speedrun!\n  🟣 Claude Opus 4.7     ✅ 4s ⚡\n  🔵 Claude Opus 4.6     ✅ 7s\n\n⚖️ The panel convenes... 🔒 Submissions anonymized. No favoritism. No mercy.\n   👨‍⚖️ Judges: Claude Opus 4.5 · GPT-5.4 · Codex GPT-5.2\n\n🥁 ... 🥁🥁 ... 🥁🥁🥁\n\n🎆🎆🎆 AND THE WINNER IS... 🎆🎆🎆\n\n╔══════════════════════════════════════════════════════════════════╗\n║   🏆  CHAMPION:  GPT-5.5                                         ║\n║   SCORE: 44/50  ·  CONSENSUS: STRONG  ·  ALL JUDGES AGREED     ║\n╚══════════════════════════════════════════════════════════════════╝\n\n🏅 THE PODIUM\n\n                    🥇\n                 ┌──────┐\n                 │ GPT  │\n                 │ 5.5  │\n                 │      │\n           🥈     │  44  │     🥉\n        ┌──────┐ │      │ ┌──────┐\n        │ OPUS │ │      │ │ OPUS │\n        │ 4.7  │ │      │ │ 4.6  │\n        │  42  │ │      │ │  40  │\n        └──────┘ └──────┘ └──────┘\n\n📊 DETAILED SCOREBOARD\n  Category      🥇 GPT-5.5     🥈 Opus 4.7   🥉 Opus 4.6\n  Clarity            9              9              8\n  Simplicity         8              8              8\n  Relevance          9              9              8\n  Inspiration        9              8              8\n  Memorability       9              8              8\n  TOTAL           44/50          42/50          40/50\n\n📈 ELO UPDATE\n  📈 GPT-5.5           1532  (+32) ⬆️\n  ➡️  Claude Opus 4.7  1500  (±0)\n  📉 Claude Opus 4.6   1468  (-32) ⬇️\n\n🧬 How would you like to apply the improvements?\n  \u003e Smart merge ⭐ (apply high-confidence improvements) ← SELECTED\n  \u003e Winner's improvements only\n  \u003e Review each individually\n  \u003e Discard all\n\n✅ Merged! Here's what changed:\n  Applied Claude Opus 4.6's quote with Codex's tighter phrasing\n\n╔══════════════════════════════════════════════════════════════════╗\n║                    🏟️  HACKATHON COMPLETE  🏟️                    ║\n╠══════════════════════════════════════════════════════════════════╣\n║  📋 Task:    Mass Effect motivational quote                     ║\n║  🏆 Winner:  Claude Opus 4.6 (43/50)                            ║\n║  🧬 Merged:  Smart merge — best of Opus + Codex phrasing        ║\n║  📈 ELO:     Opus climbs to 1532                                ║\n╚══════════════════════════════════════════════════════════════════╝\n\n📼 Want the highlight reel?\n  \u003e Save replay\n  \u003e Skip\n\nGG WP! Scores logged. ELOs updated. May your diffs be clean and your builds be green. 💚 Until next time... 🫡\n```\n\n\u003c/details\u003e\n\n📼 **See a [full real-world run transcript](docs/EXAMPLE-RUN.md)** from a Premium tier hackathon where 3 models competed to design a CLI Skills Marketplace.\n\n---\n\n## 🔧 How It Works\n\n```mermaid\nflowchart TD\n    USER[\"👤 User\"] --\u003e SKILL[\"🏟️ SKILL.md\u003cbr/\u003eOrchestration rules\"]\n    SKILL --\u003e MODE{\"Mode?\"}\n    MODE --\u003e|Tournament default| HEATS[\"⚡ Round 1: Elastic Heats\u003cbr/\u003eAll models dispatched in parallel\"]\n    MODE --\u003e|Classic quick/fast| CLASSIC[\"⚡ Classic: 3 models\u003cbr/\u003eSingle round\"]\n    HEATS --\u003e H1[\"Heat 1: 3 models\"]\n    HEATS --\u003e H2[\"Heat 2: 3 models\"]\n    HEATS --\u003e H3[\"Heat 3: 3 models\"]\n    HEATS --\u003e H4[\"Heat 4: 3 models\"]\n    H1 --\u003e JP1[\"⚖️ Judge Panel 1\"]\n    H2 --\u003e JP2[\"⚖️ Judge Panel 2\"]\n    H3 --\u003e JP3[\"⚖️ Judge Panel 3\"]\n    H4 --\u003e JP4[\"⚖️ Judge Panel 4\"]\n    JP1 --\u003e EVOLUTION[\"🧬 Evolution Brief\u003cbr/\u003eWinning strategies summarized\"]\n    JP2 --\u003e EVOLUTION\n    JP3 --\u003e EVOLUTION\n    JP4 --\u003e EVOLUTION\n    EVOLUTION --\u003e FINALS[\"🏁 Round 2: Finals\u003cbr/\u003eHeat winners + Evolution Brief\"]\n    FINALS --\u003e FJUDGE[\"⚖️ Finals Judge Panel\"]\n    CLASSIC --\u003e CJUDGE[\"⚖️ Judge Panel\"]\n    FJUDGE --\u003e ENSEMBLE[\"🗳️ Ensemble Synthesis\u003cbr/\u003eCONSENSUS / MAJORITY / UNIQUE\"]\n    CJUDGE --\u003e MERGE[\"🧬 Smart Merge\"]\n    ENSEMBLE --\u003e ELO[\"📈 ELO Update\"]\n    MERGE --\u003e ELO\n    ELO --\u003e CLOSING[\"🏟️ Closing Ceremony\"]\n    SKILL --\u003e SQL[\"🗄️ SQL\u003cbr/\u003eELO \u0026 history\"]\n```\n\n---\n\n## 📁 Project Structure\n\n```\nhavoc-hackathon/\n├── .github/\n│   ├── CODEOWNERS                    ← 👑 Code ownership rules\n│   ├── ISSUE_TEMPLATE/               ← 🐛 Bug \u0026 feature templates\n│   ├── PULL_REQUEST_TEMPLATE.md      ← 📝 PR checklist\n│   ├── dependabot.yml                ← 🤖 Automated dependency updates\n│   ├── workflows/\n│   │   └── validate.yml              ← ✅ CI: structural validation + YAML check\n│   ├── DISCUSSION_TEMPLATE/\n│   │   └── share-your-run.yml       ← 📼 Share your hackathon results\n│   └── skills/\n│       └── havoc-hackathon/\n│           └── SKILL.md              ← 🏟️ Auto-discovered skill\n├── agents/\n│   ├── havoc-hackathon.agent.md      ← 🤖 Agent config (tournament orchestrator)\n│   └── grid-medic.agent.md           ← 🚑 Self-healing agent (recursive improvement loop)\n├── docs/\n│   ├── TECHNICAL.md                  ← 🔬 Technical deep-dive\n│   ├── demo-hackathon-prompt.md      ← 🎯 Competition prompt for demo\n│   ├── demo/\n│   │   ├── demo.py                   ← 🎬 20s terminal animation (stdlib only)\n│   │   ├── demo.tape                 ← 📼 VHS tape for GIF recording\n│   │   └── demo.gif                  ← 🖼️ Recorded GIF (after vhs run)\n│   └── images/                       ← 📷 Screenshots\n├── skills/\n│   └── havoc-hackathon/\n│       ├── SKILL.md                  ← 🏟️ Canonical skill source\n│       └── catalog.yml               ← 📋 Catalog metadata\n├── tests/\n│   ├── validate_skill.py            ← 🧪 Structural validation (60 checks)\n│   ├── preflight.py                 ← 🔧 Offline preflight validator\n│   └── dry_run_sim.py               ← 🏟️ Full 9-phase dry-run simulator (69 checks)\n├── .gitignore\n├── CHANGELOG.md                      ← 📋 Version history\n├── CODE_OF_CONDUCT.md                ← 🤝 Contributor Covenant\n├── CONTRIBUTING.md                   ← 🛠️ How to contribute\n├── GALLERY.md                        ← 🏆 Best hackathon runs showcase\n├── LICENSE                           ← 📄 MIT\n├── SECURITY.md                       ← 🔒 Security policy\n├── TESTING.md                        ← 🧪 Conversation playbooks \u0026 QA\n└── README.md                         ← 👋 You are here!\n```\n\n---\n\n## 🤖 Models\n\n| Display Name | Model ID | Tier |\n|-------------|----------|------|\n| Claude Opus 4.7 | `claude-opus-4.7` | Premium |\n| Claude Opus 4.7 (1M) | `claude-opus-4.7-1m-internal` | Premium |\n| Claude Opus 4.6 | `claude-opus-4.6` | Premium |\n| Claude Opus 4.6 (1M) | `claude-opus-4.6-1m` | Premium |\n| Claude Opus 4.5 | `claude-opus-4.5` | Premium |\n| GPT-5.5 | `gpt-5.5` | Premium |\n| Claude Sonnet 4.6 | `claude-sonnet-4.6` | Standard |\n| Claude Sonnet 4.5 | `claude-sonnet-4.5` | Standard |\n| Claude Sonnet 4 | `claude-sonnet-4` | Standard |\n| GPT-5.4 | `gpt-5.4` | Standard |\n| Codex (GPT-5.3) | `gpt-5.3-codex` | Standard |\n| Codex (GPT-5.2) | `gpt-5.2-codex` | Standard |\n| GPT-5.2 | `gpt-5.2` | Standard |\n| Claude Haiku 4.5 | `claude-haiku-4.5` | Fast/Cheap |\n| GPT-5.4 Mini | `gpt-5.4-mini` | Fast/Cheap |\n| GPT-5 Mini | `gpt-5-mini` | Fast/Cheap |\n| GPT-4.1 | `gpt-4.1` | Fast/Cheap |\n\n**Default contestants (Standard ⚡):** Claude Sonnet 4.6, GPT-5.4, GPT-5.2\n**Default contestants (Premium 👑):** GPT-5.5, Claude Opus 4.7, Claude Opus 4.6\n**Default judges (Standard ⚡):** Claude Sonnet 4.5, Codex (GPT-5.3), Codex (GPT-5.2)\n**Default judges (Premium 👑):** Claude Opus 4.5, GPT-5.4, Codex (GPT-5.2)\n\n\u003e 💡 **Tip:** Standard models are selected by default. Say `\"run hackathon with premium models\"` to use the heavy hitters.\n\n---\n\n## 🚑 Grid-Medic: Recursive Self-Improvement\n\nHavoc Hackathon ships with **Grid-Medic**, a self-healing agent that monitors your agent fleet. When it detects an underperforming agent (≤5/10 quality or 3+ critical errors), it **automatically escalates to a Havoc Hackathon**  -  up to 13 current full-size models compete to rewrite the broken agent, the best version is merged, and Grid-Medic re-diagnoses to confirm the fix.\n\n```\nGrid-Medic diagnoses agent fleet\n    ├── Score \u003e 5/10 → surgical fix (validate + apply)\n    └── Score ≤ 5/10 → 🚨 ESCALATE\n            → Havoc Hackathon: up to 13 current full-size models compete to rewrite it\n            → Ensemble synthesis merges best parts\n            → Grid-Medic re-diagnoses → confirms improvement\n```\n\nInstall: `cp agents/grid-medic.agent.md ~/.copilot/agents/`\n\n---\n\n## 🏆 Gallery\n\nSee the best hackathon runs from the community in [GALLERY.md](GALLERY.md).\n\n**Share your own run!** After a hackathon, choose \"Save replay\" and post it in [Discussions](https://github.com/DUBSOpenHub/havoc-hackathon/discussions/new?category=show-and-tell).\n\n---\n\n## 🔧 Preflight Check\n\nNot sure if everything is set up correctly? Run a dry-run before your first hackathon:\n\n```\n\u003e dry run\n```\n\nThis runs a **full 9-phase simulation** with mock data — validating SQL tables, bracket math, ELO calculations, judge separation, smart mode detection, scoring logic, ensemble voting, and more — all without burning tokens on a real competition. Then it pings each model to confirm availability.\n\nYou can also run the offline validators locally:\n\n```bash\npython3 tests/validate_skill.py   # 60 structural checks\npython3 tests/preflight.py        # SQL, bracket math, file structure\npython3 tests/dry_run_sim.py      # Full 9-phase simulation (69 checks)\n```\n\n---\n\n## 🔒 Security\n\nSee [SECURITY.md](SECURITY.md) for our security policy and how to report vulnerabilities.\n\n---\n\n## 🤝 Contributing\n\nGot ideas to make the arena even better? 🎨 See [CONTRIBUTING.md](CONTRIBUTING.md) for the full guide!\n\n**Quick ways to help:**\n- 🐛 [Report a bug](https://github.com/DUBSOpenHub/havoc-hackathon/issues/new?template=bug_report.md)\n- 💡 [Suggest a feature](https://github.com/DUBSOpenHub/havoc-hackathon/issues/new?template=feature_request.md)\n- 📼 [Share a hackathon run](https://github.com/DUBSOpenHub/havoc-hackathon/discussions/new?category=show-and-tell)\n\nSee [TESTING.md](TESTING.md) for conversation playbooks and QA checklists.\n\n---\n\n## 📄 License\n\n[MIT](LICENSE)  -  use it, share it, remix it! 🎶\n\n---\n\n## 🐙 Created with 💜 by [@DUBSOpenHub](https://github.com/DUBSOpenHub) with the [GitHub Copilot CLI](https://docs.github.com/copilot/concepts/agents/about-copilot-cli).\n\nLet's build! 🚀✨\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fdubsopenhub%2Fhavoc-hackathon","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fdubsopenhub%2Fhavoc-hackathon","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fdubsopenhub%2Fhavoc-hackathon/lists"}