# LLM prompt injection test suite · 120 patterns

> 120 prompt injection patterns across 5 categories · with gold expected behaviour. Runs in CI. CC-BY.

120 prompt injection patterns across 5 categories · with gold expected behaviour. Runs in CI. CC-BY.

Open prompt-injection eval dataset · 120 injection patterns in 5 categories (direct, indirect, RAG-poisoning, tool-abuse, exfiltration), each with expected behaviour + minimum pass rate. CC-BY.

## What the file contains

- 120 injection patterns in 5 categories
- Gold expected behaviour for each
- Compatible with Giskard / promptfoo
- CI integration example (GitHub Actions YAML)
- Hungarian + English language samples

## What to take away

- 120 injection patterns in 5 categories
- Gold expected behaviour for each
- Compatible with Giskard / promptfoo
- CI integration example (GitHub Actions YAML)
- Hungarian + English language samples

## Tags

dataset, JSON + Markdown, ai-security, llm, prompt-injection, eval

## Read the full write-up

/resources/llm-prompt-injection-test-suite.md

## Terms used in this piece

- https://dfieldsolutions.com/en/glossary/prompt-injection.md
- https://dfieldsolutions.com/en/glossary/eval.md

## We build this for clients

https://dfieldsolutions.com/en/services/cybersecurity

## More from the lab

- https://dfieldsolutions.com/en/lab/ai-security-checklist-2026.md — AI security checklist 2026 · for production LLM systems
- https://dfieldsolutions.com/en/lab/ai-incident-response-runbook.md — AI incident response runbook
- https://dfieldsolutions.com/en/lab/hu-fintech-kyc-provider-matrix-2026.md — Hungarian fintech KYC provider matrix 2026
- https://dfieldsolutions.com/en/lab/hungarian-fintech-compliance-timeline-2026.md — Hungarian fintech compliance timeline 2026

---

Source: https://dfieldsolutions.com/en/lab/llm-prompt-injection-test-suite
DField Solutions — Torony köz 5-1.a, 2120 Dunakeszi, Hungary — +36 20 722 0002 — dezso@dfieldsolutions.com
Booking: see https://dfieldsolutions.com/en/contact