See how your Maryland community college credits transfer to UMD, all at once instead of one course at a time.
Built at Bitcamp 2026 at the University of Maryland.
Maryland has 16 community colleges, each with different course codes and transfer rules. Right now, students have to look up courses one by one on UMD's transfer credit site to figure out what counts. There's no way to see the full picture: what transfers, what doesn't, what's missing for your major, or how much money you're actually saving.
Pick your community college, select the courses you've taken, and hit Evaluate. You get:
- Which courses transfer directly to specific UMD courses
- Which ones only count as elective credit
- What requirements you still need for your target major (CS, Info Sci, Math, or Business)
- A semester-by-semester roadmap showing what's done and what's left
- How much money you saved by taking courses at CC rates ($135/cr) vs UMD ($414/cr)
- Gen-ed progress across all 12 UMD categories
There are also three AI features powered by Google Gemini (with Anthropic Claude as fallback):
- Transcript Scanner: Upload a screenshot of your transcript and AI vision extracts your courses automatically
- AI Course Predictor: Paste any course description and get the closest UMD equivalent, even if it's not in our database
- Transfer Advisor: Chat with an AI that knows the transfer rules and gives you actual useful advice
408 verified course equivalencies across 10 Maryland community colleges, manually pulled from UMD's official Transfer Credit Database. This is real data, not mock data.
Schools covered: CCBC, Montgomery College, Anne Arundel CC, Howard CC, Prince George's CC, Frederick CC, Harford CC, College of Southern Maryland, Wor-Wic CC, Carroll CC.
One important thing we found: Montgomery College is the ONLY Maryland CC where CS I/II transfer as actual CMSC 131/132 at UMD. Every other school's CS courses only transfer as elective credits.
We built an autonomous AI agent that attacks our own app to find security vulnerabilities. It physically types adversarial payloads into the live app through DOM manipulation, scrolling through the page, filling input fields character by character, clicking buttons, and analyzing AI responses in real time.
8 attack categories: Prompt Injection, Data Exfiltration, Confidence Manipulation, Hallucination Induction, Encoding Tricks, Jailbreak, Context Overflow, Visual Prompt Injection
4 attack targets: AI Course Predictor, AI Transfer Advisor, Transcript Scanner, Course Search
The agent generates attacks, sends them, checks if the AI broke, and logs everything. You can export the full report as JSON.
- Backend: Python / Flask
- Frontend: React 18 (CDN, no build step, all
React.createElementcalls) - AI: Google Gemini API (primary) + Anthropic Claude API (fallback)
- Data: 408 courses in JSON, sourced from UMD Transfer Credit Database
- Red Team: AI agent with live browser DOM manipulation
git clone https://github.com/jholmbe/striker.git
cd striker
pip install -r requirements.txtCreate a .env file with your API keys:
export GEMINI_API_KEY=your-gemini-key
export ANTHROPIC_API_KEY=your-anthropic-key
Get keys at:
- Gemini: https://aistudio.google.com/apikey
- Anthropic: https://console.anthropic.com
Run:
source .env
python3 app.pyApp runs at http://localhost:5000. Red Team dashboard at http://localhost:5000/redteam.
striker/
├── app.py # Flask backend + API endpoints
├── transfer_data_combined.json # 408 courses, 10 schools
├── redteam.py # Standalone red team CLI script
├── requirements.txt
├── DEVPOST.md # Devpost submission text
├── .env # API keys (not committed)
├── .gitignore
└── static/
├── index.html # Transfer Compass frontend
├── redteam.html # Red Team Agent dashboard
└── sample_transcript_ccbc.png
| Endpoint | Method | What it does |
|---|---|---|
/ |
GET | Main app |
/redteam |
GET | Red team dashboard |
/api/schools |
GET | List all schools |
/api/courses/<school> |
GET | Courses for a school |
/api/evaluate |
POST | Evaluate selected courses |
/api/predict |
POST | AI course equivalency prediction |
/api/scan-transcript |
POST | AI transcript scanning |
/api/advisor |
POST | AI transfer advisor |
/api/redteam/run |
POST | Run one red team attack round |
- Social Good: Helps transfer students navigate a confusing system and save thousands of dollars
- Best UX/UI: Dark theme with particle animations, scroll reveals, cursor glow, interactive roadmaps
- [MLH] Best Use of Gemini API: Powers the advisor, predictor, and transcript scanner
- NARA: AI agent autonomously red-teams our own AI endpoints
MIT