Hackathon Skills

Production-oriented, cross-platform Agent Skills that help AI coding agents strategically analyze, build, validate, demo, pitch, and submit hackathon projects with the strongest realistic chance of winning.

MIT Licensed Zero Dependencies Cross-Platform NPM Package
Get Started View on GitHub

Key Features

The Hackathon Grand Prize Agent Skill optimizes for one memorable, working core workflow—not feature count. It combines hackathon research, idea scoring, differentiation, architecture, sponsor strategy, execution planning, engineering gates, security review, demo recovery, judge red-teaming, and submission readiness.

◉

Strategic Analysis

Analyze hackathon rules, identify winning criteria, and align with sponsor requirements for maximum competitive advantage.

✦

Idea Scoring

Score and rank project ideas based on differentiation, feasibility, sponsor fit, and realistic winning potential.

▦

Architecture Planning

Design right-sized architectures that prove differentiators within time constraints with built-in failure handling.

⛨

Security Review

Built-in security gates, secret scanning, input validation, and privacy checks before submission.

▶

Demo Engineering

Create deterministic demos with fallback paths, fixtures, and recovery strategies for reliable presentations.

✎

Pitch Framework

Structure compelling, evidence-backed pitches optimized for judges with timing and impact considerations.

Installation

1Quick Install with NPX

Install the skill directly into your agent's project directory without adding dependencies:

Terminal
npx hackathon-skills add hackathon-grand-prize --agent claude

2List Available Skills

View the catalog of available hackathon skills:

Terminal
npx hackathon-skills list

3Advanced Options

Customize installation with additional flags:

Terminal
# Install for specific scope
npx hackathon-skills add hackathon-grand-prize --agent claude --scope user

# Dry run to preview changes
npx hackathon-skills add hackathon-grand-prize --agent cursor --dry-run

# Force reinstall with backup
npx hackathon-skills add hackathon-grand-prize --agent kiro --force

# Custom target directory
npx hackathon-skills add hackathon-grand-prize --target ./my-skills
# List supported agents
npx hackathon-skills agents

Supported AI Coding Agents

Hackathon Skills works seamlessly with all major AI coding agents that support the Agent Skills specification:

Claude Code

Anthropic's AI coding assistant

.claude/skills/hackathon-grand-prize/
Terminal
npx hackathon-skills add hackathon-grand-prize --agent claude

Cursor

AI-powered code editor

.cursor/skills/hackathon-grand-prize/
Terminal
npx hackathon-skills add hackathon-grand-prize --agent cursor

GitHub Copilot

VS Code / GitHub's AI pair programmer

.github/skills/hackathon-grand-prize/
Terminal
npx hackathon-skills add hackathon-grand-prize --agent copilot

Kiro

Next-gen AI coding assistant

.kiro/skills/hackathon-grand-prize/
Terminal
npx hackathon-skills add hackathon-grand-prize --agent kiro

OpenAI Codex

OpenAI's code generation model

.agents/skills/hackathon-grand-prize/
Terminal
npx hackathon-skills add hackathon-grand-prize --agent codex

Google Antigravity CLI

Google's AI development tool

.agents/skills/hackathon-grand-prize/
Terminal
npx hackathon-skills add hackathon-grand-prize --agent antigravity

How to Use

Natural Language Commands

Simply ask your AI agent naturally:

Analyze Hackathon

"Analyze this hackathon's official rules and recommend three differentiated ideas."

Review Project

"Review our current project like a hostile judge and fix the highest-impact gaps."

Create Demo

"Create a deterministic three-minute demo with fallback and recovery paths."

Prepare Submission

"Prepare and validate our final README, pitch, and submission checklist."

Documented Modes

Or invoke specific documented modes:

/hackathon-analyze - Intelligence report and winning surface
/hackathon-ideas - Ranked candidates and thesis
/hackathon-build - Implemented vertical slice
/hackathon-review - Findings and remediation
/hackathon-demo - Rehearsable demo runbook
/hackathon-pitch - Timed, evidence-backed script
/hackathon-submit - Verified submission package
/hackathon-score - Evidence-backed scoring

Skill Capabilities

The Hackathon Grand Prize Strategist provides comprehensive support through multiple execution phases:

Phase 0: Research

Verify hackathon rules, eligibility, restrictions, judging criteria, and submission requirements with dated official citations.

Phases 1-4: Ideation

Discover ideas, score candidates, differentiate from competition, and form compelling product thesis.

Phases 5-7: Planning

Design architecture, control MVP scope, and engineer deterministic demo paths first.

Phases 8-10: Quality

Raise UX quality, prove agent quality, and run comprehensive engineering gates.

Phases 11-12: Validation

Attack as a judge, identify weak answers, score readiness with evidence-backed rationale.

Phases 13-18: Delivery

Plan execution, build pitch and demo scripts, assemble submission package, final red team review.

Design Principles

  • One canonical, self-contained skill package - Everything you need in one place
  • Trigger-rich metadata and explicit mode routing - Clear invocation patterns
  • Progressive disclosure through focused references - Load only what you need
  • Official-source research with dated citations - No invented facts
  • Evidence-based decision making - Facts, assumptions, and proof strength tracked separately
  • Human approval for critical actions - Safety gates for destructive operations
  • Deterministic validation and demo fallbacks - Reliability built in

Validation & Helper Scripts

Optional Python 3.9+ helper scripts (zero dependencies, standard library only):

Python
# Validate skill structure
python hackathon-grand-prize/scripts/validate-skill.py

# Score project readiness
python hackathon-grand-prize/scripts/score-project.py --scores-json '<scores>'

# Check project health
python hackathon-grand-prize/scripts/project-health-check.py --checks-json '<checks>'
python hackathon-grand-prize/scripts/validate-submission.py --file README.md --check-local-links