url-prompt-injection
OfficialDetect URL-based prompt injection in AI chat UIs.
System Documentation
What problem does it solve?
This skill addresses the risk of cross-site prompt injection vulnerabilities in AI chatbot, assistant, and agent web interfaces that accept user prompts via URL query parameters, which attackers can exploit to execute malicious prompts in authenticated victim sessions to steal data or perform unauthorized actions.
Core Features & Use Cases
- 5-Stage Static Analysis Workflow: Systematically scan JavaScript source code to identify vulnerable URL parameter handling, including detection of query param consumption, nuqs state declarations, and auto-submit wiring in chat components.
- Exploitability Assessment Framework: Guide testers through confirming parameter behavior, evaluating agent tool access and data exfiltration channels, and calibrating vulnerability severity based on impact.
- Red Teaming Use Case: Ideal for security teams testing internal AI copilots, customer support chatbots, or agent UIs for injection flaws that could lead to data breaches or unauthorized tool execution.
Quick Start
Use the url-prompt-injection skill to scan your target application's JavaScript source code for URL-based prompt injection vulnerabilities in its AI chat interface and generate a prioritized severity assessment.
Dependency Matrix
Required Modules
None requiredComponents
Standard package💻 Claude Code Installation
Recommended: Let Claude install automatically. Simply copy and paste the text below to Claude Code.
Please help me install this Skill: Name: url-prompt-injection Download link: https://github.com/dreadnode/capabilities/archive/main.zip#url-prompt-injection Please download this .zip file, extract it, and install it in the .claude/skills/ directory.
Agent Skills Search Helper
Install a tiny helper to your Agent, search and equip skill from 536,000+ vetted skills library on demand.