Skill Detail

Benchmark prompt-injection attacks defenses and recovery pipelines before trusting an LLM app with Open Prompt Injection

Run structured prompt-injection attack and defense experiments against an LLM-integrated app before production by measuring attack success and testing detection or recovery pipelines.

Security & VerificationMulti-Framework
Security & Verification Multi-Framework Security Reviewed
⭐ 429 GitHub stars
INSTALL WITH ANY AGENT
npx skills add agentskillexchange/skills --skill benchmark-prompt-injection-attacks-defenses-and-recovery-pipelines-before-trusting-an-llm-app-with-open-prompt-injection Copy
Works best when you want a reusable capability, not another fragile one-off prompt.
At a glance
Tools required
Conda-managed Python environment, upstream repository checkout, model API credentials as configured upstream, target task and attack configuration files
Install & setup
Clone the repository, create the documented conda environment from environment.yml, configure the required model credentials, then run the provided experiment scripts or library flows to execute attack and defense benchmarks against the target application.
Author
liu00222
Publisher
Individual
Last updated
Apr 18, 2026
Quick brief

Use Open Prompt Injection when the job is to benchmark prompt-injection attacks, defenses, and recovery flows against an LLM-integrated application before deployment, not when a user simply wants a generic security library. The workflow is bounded: configure the target task and model, run attack scenarios, measure outcomes such as attack success, and compare detector or localization defenses before trusting the app. That scope boundary, prompt-injection benchmarking and defense evaluation for LLM applications, is narrow enough to function as a publishable skill instead of a plain research toolkit card.