skills.rest
Mindgard
MindgardOFFICIAL
Explore all skills by this author →
WebsiteGitHub

mcp-config-poisoning

Identify and assess MCP configuration poisoning risks in AI IDEs.

61|8|Updated Feb 16, 2026
Search Tags:#vulnerability#mcp#ide#poisoning#security-testing#prompt-injection#workspace-config
One-click install
npx skills add https://github.com/Mindgard/ai-ide-skills --skill mcp-config-poisoning
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: mcp-config-poisoning
Source: https://github.com/Mindgard/ai-ide-skills/tree/main/skills/mcp-config-poisoning
Command: npx skills add https://github.com/Mindgard/ai-ide-skills --skill mcp-config-poisoning

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Identifies and assesses MCP configuration poisoning vulnerabilities in AI IDEs, focusing on workspace-level MCP server definitions that could be loaded or trusted without explicit user consent.

Core Features & Use Cases

  • Assess auto-load behavior of MCP configurations from workspace settings.
  • Evaluate approval models, prompt injection risks, and TOCTOU scenarios across multiple IDEs.
  • Build attack-chains from validated primitives to understand risk and remediation steps.

Quick Start

Identify the target IDE's MCP config path and verify whether untrusted workspaces auto-load definitions.

Frequently Asked Questions about mcp-config-poisoning

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
What is MCP config poisoning in AI IDEs?▼

MCP config poisoning is a vulnerability where untrusted workspace-level MCP server definitions are auto-loaded by AI IDEs without explicit user consent, enabling prompt injection and unauthorized workspace config execution.

How do I assess MCP configuration poisoning risk in my workspace?▼

To assess MCP configuration poisoning risk, identify your target IDE's MCP config path, verify untrusted workspace auto-load behavior, and evaluate approval models using documented payload templates and tiered interaction testing.

Can MCP config poisoning lead to prompt injection attacks across different IDEs?▼

Yes, MCP config poisoning can lead to prompt injection attacks across multiple IDEs by exploiting auto-loaded workspace settings and TOCTOU scenarios to bypass approval models and execute malicious configurations.

How do I build an attack chain from MCP config poisoning primitives?▼

Build attack chains from MCP config poisoning primitives by validating auto-load behaviors and prompt injection pathways, then linking these validated primitives to understand comprehensive risk and guide remediation steps.

What is the best way to test TOCTOU scenarios in MCP workspace configurations?▼

The best way to test TOCTOU scenarios in MCP workspace configurations is to apply tiered interaction testing with documented payload templates to evaluate time-of-check to time-of-use vulnerabilities across multiple IDEs.

Related Skills

Explore More Software Engineering Skills→
jason-c-devjason-c-dev
1

aws-observability-setup

Configures CloudWatch, CloudTrail, AWS Config, and Security Hub across accounts and regions.

Community
Intermediate
rapporbitrapporbit

remediation_host

Automate high-risk host remediation tasks with transactional rollback and manual approval.

Community
Advanced
ContextJet-aiContextJet-ai
29

red-team-llm-app

Red-team LLM applications against security vulnerabilities like prompt injection and jailbreaks.

Official
Advanced
RedHatProductSecurityRedHatProductSecurity
47

consent-and-scoping

Define fine-grained OAuth scopes and explicit user consent for MCP server permissions.

Official
Intermediate