security-scan

Scan Claude Code configuration files for security vulnerabilities and misconfigurations.

1|Updated Mar 6, 2026
One-click install
npx skills add https://github.com/khetansarvesh/ai_skills_repo --skill security-scan-khetansarvesh
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: security-scan
Source: https://github.com/khetansarvesh/ai_skills_repo/tree/main/skills/security-scan
Command: npx skills add https://github.com/khetansarvesh/ai_skills_repo --skill security-scan-khetansarvesh

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill helps you prevent security breaches by identifying risky patterns and misconfigurations in your Claude Code project setup before they can be exploited through prompt injection, unsafe tool access, or secret leakage.

Core Features & Use Cases

  • Configuration Security Auditing: Reviews common Claude Code configuration files for vulnerabilities, including CLAUDE.md and settings.json misconfigurations.
  • Injection and Misuse Detection: Flags prompt injection vectors and unsafe instruction patterns that can enable malicious behavior.
  • Tooling and Integration Risk Checks: Examines MCP server definitions, hooks, and agent definitions for risky execution paths and unsafe permissions.
  • Actionable Remediation: Supports safe automated fixes for auto-fixable issues to reduce manual hardening effort.

Quick Start

Run a scan of your current project’s Claude Code configuration by executing: npx ecc-agentshield scan --path .claude .

Frequently Asked Questions about security-scan

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

FAQPage Schema
How do I scan Claude Code configuration files for prompt injection risks?

Scan Claude Code configuration files for prompt injection risks by running AgentShield against the .claude directory. This process detects misconfigurations, unsafe instruction patterns, and integration vulnerabilities in CLAUDE.md and settings.json.

What security vulnerabilities can be found in CLAUDE.md and settings.json?

Security vulnerabilities in CLAUDE.md and settings.json include prompt injection vectors, unsafe tool access, and secret leakage. Scanning also identifies risky execution paths and unsafe permissions within MCP server definitions, hooks, and agent definitions.

Can I automatically fix misconfigurations found during a Claude Code security scan?

Yes, you can automatically fix misconfigurations found during a Claude Code security scan. The process supports safe auto-fixes for auto-fixable issues to reduce manual hardening effort across your configuration files.

Does the security scan work for CI/CD gating and post-edit verification?

Yes, the security scan works for CI/CD gating and post-edit verification. It applies to new project setup, periodic hygiene checks, and post-edit verification of CLAUDE.md, producing terminal, JSON, Markdown, or HTML reports.

What is the best way to harden MCP server definitions against unsafe execution paths?

The best way to harden MCP server definitions against unsafe execution paths is running a configuration security audit. This examines MCP server definitions, hooks, and agent definitions for risky execution paths, unsafe permissions, and deeper adversarial risks.

What report formats are available for Claude Code configuration security audits?

Available report formats for Claude Code configuration security audits include terminal, JSON, Markdown, and HTML. These reports detail detected vulnerabilities, misconfigurations, and prompt injection risks found within the .claude directory.