PromptIntel (NOVA)
Free collaborative database of adversarial prompts — Indicators of Prompt Compromise (IoPC).
What it does
A free collaborative database of adversarial prompts known as Indicators of Prompt Compromise (IoPC). Created by Thomas Roccia as part of the NOVA prompt pattern matching ecosystem. 38+ techniques classified across four categories: instruction manipulation, misuse, extraction, and evasion. API available for integration.
Security relevance
The first community-driven threat intelligence feed specifically for adversarial prompts — think VirusTotal for prompt attacks. Each entry includes the prompt, classification, technique mapping, and metadata. Searchable by tag, technique, and keyword. Integrates with the NOVA detection engine for automated prompt pattern matching.
When to use it
Use for red teaming research, building detection rules, and testing guardrail effectiveness against known adversarial prompts. The API enables automated integration into security workflows. Contribute your own findings to strengthen the community database.
OWASP coverage
Risks addressed — mapped to both OWASP Top 10 standards. 3 in LLM, 1 in Agentic.
The raw record
What Yuntona stores. Single source of truth — fork it on GitHub.
name: PromptIntel (NOVA) slug: promptintel-nova type: Mixed category: AI Red Teaming url: https://promptintel.novahunting.ai reviewed: 2026-04 added: 2026-04 updated: 2026-04 risks: llm: [LLM01, LLM02, LLM06] asi: [ASI01] complexity: Plug & Play pricing: — audience: Red Team lifecycle: [monitor] tags: [Database, Injection, Open Source, Threat Intel]