# ThePromptDefence

- **Event:** [RAISE Summit Hackathon](https://cerebralvalley.ai/e/raise-summit-hackathon)
- **When:** Jul 4 at 9:00 AM – Jul 5 at 7:00 PM (GMT+2)
- **Where:** Paris, France
- **Team:** [Aniket Karne](https://cerebralvalley.ai/u/aniketkarne)
- **GitHub:** https://github.com/aniketkarne/ThePromptDefence
- **Gallery:** https://cerebralvalley.ai/e/raise-summit-hackathon/hackathon/gallery
- **Page:** https://cerebralvalley.ai/e/raise-summit-hackathon/hackathon/gallery/94

The youtube video is not needed since this is a ThePromptDefence is a local-first, MCP-native prompt injection firewall for Cursor. It runs a 3-layer detection pipeline — regex heuristics, DeBERTa-v3 semantic classifier, and structural obfuscation analysis — entirely on-device, with ~29ms p99 latency and 0% false positives. As an MCP server, it plugs into Cursor in 2 minutes and intercepts malicious prompts before they reach Claude or GPT-4. Zero API cost, zero data egress, fully offline-capable.

## More from RAISE Summit Hackathon

- [Weave](https://cerebralvalley.ai/e/raise-summit-hackathon/hackathon/gallery/91)
- [AlphaGo](https://cerebralvalley.ai/e/raise-summit-hackathon/hackathon/gallery/92)
- [MOE](https://cerebralvalley.ai/e/raise-summit-hackathon/hackathon/gallery/93)
- [Roboss](https://cerebralvalley.ai/e/raise-summit-hackathon/hackathon/gallery/95)
- [eHoReCa!](https://cerebralvalley.ai/e/raise-summit-hackathon/hackathon/gallery/96)
- [SelfAware](https://cerebralvalley.ai/e/raise-summit-hackathon/hackathon/gallery/97)

---

Markdown version of https://cerebralvalley.ai/e/raise-summit-hackathon/hackathon/gallery/94. Site index for agents: https://cerebralvalley.ai/llms.txt · full text: https://cerebralvalley.ai/llms-full.txt