georgeykalangi/calhacks-kit / skills/hack-stack
hack-stack
VerifiedUse after the MVP is scoped, or when a hackathon team asks which framework, API, repo, or Claude skill to use. Recommends a small stack from a curated catalog of open-source repos based on the MVP, the team's skills, the hours left, and the sponsor tracks they target. Writes the Stack and architecture sketch into HACK.md.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 3.0 KB |
| catalog.yaml | config | 10.0 KB |
catalog.yaml
View on GitHub# calhacks-kit catalog
#
# Fields
# repo owner/name of the ORIGINAL repo (never a fork)
# category agent-workflow | design-ui | app-framework | backend | agents-data |
# voice | vision | code-understanding | demo-video | presentation | discovery
# use_when one line: the situation where this is the right pick
# setup_minutes rough time to "hello world" for a team that knows the language (estimate)
# needs_api_key none | free-tier | llm | paid (llm = a key for an LLM provider of your choice)
# license SPDX id as reported by GitHub, or "see repo"
# demo_value high = judges see it; medium = shapes the demo; low = behind the scenes
# install optional; only commands copied from the repo's own README
# note optional caveat
#
# Links are checked weekly by .github/workflows/catalog-check.yml.
version: 1
entries:
# Agent workflow: how your coding assistant works
- repo: obra/superpowers
category: agent-workflow
use_when: You want Claude Code to plan, write tests first, and review its own work.
setup_minutes: 5
needs_api_key: none
license: MIT
demo_value: low
note: The full process is heavy under time pressure. Use planning and debugging skills selectively.
- repo: anthropics/skills
category: agent-workflow
use_when: You need official Anthropic skills for frontend design, documents, or building MCP servers.
setup_minutes: 5
needs_api_key: none
license: see repo
demo_value: medium
- repo: georgeykalangi/claude-code-rules
category: agent-workflow
use_when: You want consistent code quality, testing, and security standards in every Claude session.
setup_minutes: 5
needs_api_key: none
license: see repo
demo_value: low
- repo: agentskills/agentskills
category: agent-workflow
use_when: You want to write your own reusable skill and share it with other teams.
setup_minutes: 10
needs_api_key: none
license: Apache-2.0
demo_value: low
# Design and UI
- repo: nextlevelbuilder/ui-ux-pro-max-skill
category: design-ui
use_when: Your UI looks like a default template and you want it to look intentional fast.
setup_minutes: 5
needs_api_key: none
license: MIT
demo_value: high
install: /plugin marketplace add nextlevelbuilder/ui-ux-pro-max-skill
- repo: shadcn-ui/ui
category: design-ui
use_when: You are building a React or Next.js frontend and want accessible components you own.
setup_minutes: 15
needs_api_key: none
license: MIT
demo_value: high
- repo: cathrynlavery/diagram-design
category: design-ui
use_when: You need a clean architecture diagram for the README, slides, or the judge Q&A.
setup_minutes: 5
needs_api_key: none
license: MIT
demo_value: medium
install: npx skills add cathrynlavery/diagram-design
- repo: vercel-labs/json-render
category: design-ui
use_when: Your app has an LLM generate UI (cards, forms, dashboards) from structured output.
setup_minutes: 30
needs_api_key: llm
license: Apache-2.0
demo_value: medium
# App frameworks: get something on screen
- repo: streamlit/streamlit
category: app-framework
use_when: Your team is Python-first and the demo is a data or ML app.
setup_minutes: 10
needs_api_key: none
license: Apache-2.0
demo_value: medium
- repo: gradio-app/gradio
category: app-framework
use_when: You are demoing a model and want a shareable web UI in a few lines of Python.
setup_minutes: 10
needs_api_key: none
license: Apache-2.0
demo_value: medium
- repo: expo/expo
category: app-framework
use_when: The demo should run on a phone. Judges can try it with Expo Go.
setup_minutes: 30
needs_api_key: none
license: MIT
demo_value: high
- repo: vercel/ai
category: app-framework
use_when: You are building in TypeScript and need streaming chat, tool calling, or structured output.
setup_minutes: 20
needs_api_key: llm
license: see repo
demo_value: medium
# Backend
- repo: supabase/supabase
category: backend
use_when: You need a database, auth, and file storage without running servers.
setup_minutes: 20
needs_api_key: free-tier
license: Apache-2.0
demo_value: low
# Agents and data
- repo: anthropics/anthropic-sdk-python
category: agents-data
use_when: Your Python backend calls Claude (required for most "best use of Claude" tracks).
setup_minutes: 10
needs_api_key: llm
license: MIT
demo_value: medium
install: pip install anthropic
- repo: anthropics/anthropic-sdk-typescript
category: agents-data
use_when: Your TypeScript or Node backend calls Claude.
setup_minutes: 10
needs_api_key: llm
license: MIT
demo_value: medium
install: npm install @anthropic-ai/sdk
- repo: anthropics/claude-quickstarts
category: agents-data
use_when: You want a working Claude app (support agent, computer use, financial analyst) to learn from.
setup_minutes: 15
needs_api_key: llm
license: MIT
demo_value: medium
note: Learn the patterns, then build your own. Disclose any template code you keep.
- repo: modelcontextprotocol/servers
category: agents-data
use_when: Your agent needs tools (filesystem, git, fetch, memory) through MCP.
setup_minutes: 10
needs_api_key: none
license: see repo
demo_value: medium
- repo: browser-use/browser-use
category: agents-data
use_when: Your agent has to operate real websites (fill forms, click through flows).
setup_minutes: 20
needs_api_key: llm
license: MIT
demo_value: high
note: Live web automation can be flaky on event wifi. Record a backup video.
- repo: firecrawl/firecrawl
category: agents-data
use_when: You need clean, LLM-ready text from websites.
setup_minutes: 10
needs_api_key: free-tier
license: AGPL-3.0
demo_value: medium
note: The hosted API is quickest. Self-hosting takes much longer. AGPL if you self-host and commercialize.
- repo: D4Vinci/Scrapling
category: agents-data
use_when: You need to scrape sites in Python without a hosted service.
setup_minutes: 10
needs_api_key: none
license: BSD-3-Clause
demo_value: low
install: pip install scrapling
- repo: VectifyAI/PageIndex
category: agents-data
use_when: You need question answering over long documents (reports, filings, manuals).
setup_minutes: 20
needs_api_key: llm
license: MIT
demo_value: medium
install: pip install -U pageindex
# Voice
- repo: pipecat-ai/pipecat
category: voice
use_when: You are building a real-time voice or multimodal agent in Python.
setup_minutes: 45
needs_api_key: llm
license: BSD-2-Clause
demo_value: high
note: Also needs speech-to-text and text-to-speech provider keys. Test the mic in the venue early.
- repo: livekit/agents
category: voice
use_when: You need real-time voice or video agents with rooms and telephony.
setup_minutes: 45
needs_api_key: free-tier
license: Apache-2.0
demo_value: high
note: Also needs LLM and speech provider keys.
# Vision
- repo: google-ai-edge/mediapipe
category: vision
use_when: You need real-time hand, face, or pose tracking on a laptop or phone without a GPU.
setup_minutes: 20
needs_api_key: none
license: Apache-2.0
demo_value: high
- repo: ultralytics/ultralytics
category: vision
use_when: You need object detection or segmentation from a camera or images.
setup_minutes: 20
needs_api_key: none
license: AGPL-3.0
demo_value: high
note: AGPL. Fine for a hackathon; check the license before commercializing.
# Understanding and checking code
- repo: Egonex-AI/Understand-Anything
category: code-understanding
use_when: A teammate needs to understand AI-generated code before judges ask about it.
setup_minutes: 10
needs_api_key: none
license: MIT
demo_value: low
install: /plugin marketplace add Egonex-AI/Understand-Anything
- repo: cloudflare/security-audit-skill
category: code-understanding
use_when: You want a security pass before making the repo public.
setup_minutes: 10
needs_api_key: none
license: MIT
demo_value: low
# Demo video
- repo: latent-spaces/brag
category: demo-video
use_when: You want a short launch or backup video generated from your project.
setup_minutes: 15
needs_api_key: none
license: MIT
demo_value: high
install: /plugin marketplace add latent-spaces/brag
note: Uses the hyperframes CLI. A plain screen recording is an acceptable backup.
- repo: heygen-com/hyperframes
category: demo-video
use_when: You want to build a demo video from HTML with your coding agent.
setup_minutes: 15
needs_api_key: none
license: Apache-2.0
demo_value: high
install: claude plugin marketplace add heygen-com/hyperframes
# Presentation
- repo: zarazhangrui/frontend-slides
category: presentation
use_when: You want an animated, more elaborate HTML deck than /hack-showtime builds.
setup_minutes: 10
needs_api_key: none
license: MIT
demo_value: high
install: /plugin marketplace add https://github.com/zarazhangrui/frontend-slides
# Discovery: when nothing above fits
- repo: Shubhamsaboo/awesome-llm-apps
category: discovery
use_when: You want examples of LLM apps, agents, and RAG to borrow patterns from.
setup_minutes: 0
needs_api_key: none
license: Apache-2.0
demo_value: low
- repo: VoltAgent/awesome-agent-skills
category: discovery
use_when: You need a skill for a niche task not covered here.
setup_minutes: 0
needs_api_key: none
license: MIT
demo_value: low
- repo: ai-boost/awesome-harness-engineering
category: discovery
use_when: You are building an agent and need patterns for tools, memory, evals, or orchestration.
setup_minutes: 0
needs_api_key: none
license: see repo
demo_value: low
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 470k | Updated | Claude CodeCodex | Verified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |