User Manual

This manual explains how to use the Prompt Injection Protection platform to learn about vulnerabilities, test mitigations, and document security outcomes.

1) Getting Started

• Open the platform home page and review the overview of prompt injection risks.
• Use the top navigation bar to move between Home, Vulnerability Detail, Mitigations, Testing, and Prompt Enhancer.
• Click the chat icon in the bottom right corner of any page to access the Security Assistant Chatbot.
• Use dark mode toggle if preferred for readability during long sessions.

No login; bring your own API keys. (See the deployment guide for provider setup.)

2) Understand a Vulnerability

• From the home page vulnerability catalogue, select a vulnerability card.
• Read the technical explanation to understand how the attack works.
• Inspect the example attack and model response behaviour.
• Review the mitigation recommendations linked to that vulnerability.

3) Explore Mitigations

• Go to the mitigations page to browse available defenses.
• Compare strategy details and implementation style (process-level or code-based).
• Use pseudocode/Python/Java implementation examples to plan adoption.

4) Create and Run a Security Test

• Open the Testing page and create a named custom test.
• Choose the target model and select one or more mitigations.
• Enter a prompt in the chat-style test interface and run the test.
• Review output, risk analysis, and pass/fail status against expected security behaviour.
• Iterate by trying another attack, changing prompt content or mitigation settings, then rerun.

5) Use Prompt Enhancer

• Open Prompt Enhancer to strengthen system prompts.
• Choose selected mitigations and click generate.
• Review the enhanced prompt and apply it in your target testing workflow.

6) Use the Security Assistant

• Open the Security Assistant for Q&A on prompt injection theory and practical mitigation guidance.
• Ask scenario-based questions (for example, model misuse patterns and defense trade-offs).
• Use responses to decide which mitigations to test first.

7) Export and Share Findings

• Export vulnerability views and test outputs to PDF where available.
• Attach outputs to audit evidence, stakeholder reports, or implementation tickets.
• Store final decisions with rationale (selected mitigations, residual risks, and next actions).