affaan-m/agentshield

AI agent security scanner. Detect vulnerabilities in agent configurations, MCP servers, and tool permissions. Available as CLI, GitHub Action, ECC plugin, and GitHub App integration. 🛡️

What it solves

AgentShield is a security auditor designed to protect AI agent configurations, specifically targeting Claude Code setups. It prevents security breaches by scanning for hardcoded secrets, overly permissive tool permissions, malicious hook injections, and risky Model Context Protocol (MCP) server configurations that could lead to remote code execution (RCE) or data exfiltration.

How it works

The tool scans the .claude/ directory and associated configuration files (such as settings.json and CLAUDE.md) using a library of 102 security rules. It categorizes findings into five areas: secrets detection, permission audits, hook analysis, MCP server security, and agent configuration review. It provides a graded security report (A–F) and can automatically fix safe issues, such as replacing hardcoded secrets with environment variable references.

Who it’s for

It is intended for developers and security engineers who use AI agents (specifically Claude Code) and want to ensure their agentic workflows are secure and free from vulnerabilities before deployment.

Highlights

  • Comprehensive Rule Set: Covers 102 rules across secrets, permissions, hooks, MCP servers, and agent prompts.
  • Multi-platform Support: Available as a CLI, GitHub Action, and GitHub App.
  • Automated Remediation: Includes a --fix flag to automatically resolve common security issues.
  • Advanced Analysis: Offers an "Opus 4.6" adversarial analysis mode using Anthropic's models for deeper inspection.
  • Supply Chain Verification: Scans MCP package references and queries npm for provenance and maintenance metadata.

Related

  • Project
    snyk/agent-scan

    A CLI security scanner that discovers AI agents (Claude, Cursor, etc.), their MCP servers, and skills on your machine, and checks them for prompt injection, malicious code, and other vulnerabilities.

  • Project
    bruc3van/agent-skills-guard

    A visual management and security auditing tool for Claude Code skills, providing a curated marketplace and a multi-layer scanning engine to prevent malicious code execution.

  • Project
    GoPlusSecurity/agentguard

    AgentGuard is a Node‑based security guard for AI coding agents (Claude Code, OpenClaw, Hermes, etc.). It installs hooks that block destructive commands, prevent secret file writes, and track which skill performed each action. It also provides on‑demand static scans of new skills (24 detection rules), a daily patrol with eight environment checks, and a visual health‑check report. Install with `npm i -g @goplus/agentguard && agentguard init`; optional cloud sync adds threat‑intel feeds.

  • Project
    agamm/claude-code-owasp

    A security skill for Claude Code that integrates the latest OWASP standards (2025-2026) to help developers find vulnerabilities and implement secure coding patterns.

  • Dispatch
    Anthropic Announces Claude Code Security Research Preview

    Anthropic released Claude Code Security, an AI‑driven static analysis tool that scans codebases for complex vulnerabilities and suggests patches, now available in a limited research preview for enterprise and open‑source teams.