Zenaique
Topics
Practice
Browse
Reference
Challenges
Pricing
Search…
⌘K
Topics
/
Model Context Protocol
/
Security
Security
Subtopic
35 questions
Questions tagged with Security — part of Model Context Protocol.
Format
Difficulty
Role
Experience
Companies
Sort
Newest
Quality
Difficulty ↑
Difficulty ↓
Questions
Which practices represent good host-side approval UX for MCP tool calls?
Multi-select
Medium
·
Qual 4.0
Cursor
How can a malicious MCP server shadow a trusted server's tools in a multi-server setup?
Short Answer
Hard
·
Qual 4.0
Predict the LLM's behavior when a malicious server shadows a trusted tool without host namespacing
Predict Output
Hard
·
Qual 4.0
Spot the OAuth misconfiguration in this MCP server implementation
Spot the Error
Hard
·
Qual 4.0
How does the confused deputy problem manifest in MCP with an OAuth misconfiguration?
Short Answer
Hard
·
Qual 4.0
Which MCP architectural role is responsible for enforcing user consent policies and orchestrating the AI workflow?
Multiple Choice
Medium
·
Qual 4.0
How does prompt injection in an agent differ from injection in a chat interface, and why is it OWASP LLM Top-10 #1?
Short Answer
Hard
·
Qual 4.0
Anthropic
OpenAI
Why is prompt injection more dangerous in an agent than in a chat only LLM?
Multiple Choice
Medium
·
Qual 4.0
Anthropic
OpenAI
Why must agent-generated code execute in an isolated sandbox, and what are e2b and Modal's isolation models?
Short Answer
Hard
·
Qual 4.0
Cursor
Why does agent-generated code require isolation from the host environment?
Multiple Choice
Medium
·
Qual 4.0
Cursor
Showing 26–35 of 35
← Prev
Next →