Skill Detail
Scan Claude Code configs for secrets permission drift and unsafe MCP hookups with AgentShield
Audit a Claude Code setup before use by flagging hardcoded secrets, broad allow rules, risky hooks, and dangerous MCP server config.
Security & VerificationClaude Code
Security & Verification
Claude Code
Published
⭐ 388 GitHub stars
INSTALL WITH ANY AGENT
npx skills add agentskillexchange/skills --skill scan-claude-code-configs-for-secrets-permission-drift-and-unsafe-mcp-hookups-with-agentshield
Works best when you want a reusable capability, not another fragile one-off prompt.
At a glance
Tools required
Claude Code configuration directory, AgentShield CLI or npx path, local shell access, optional CI environment for GitHub Action usage
Install & setup
Run AgentShield with the no-install npx flow or install the documented package globally, then scan the target Claude Code configuration directory and review or apply the reported fixes.
Author
affaan-m
Publisher
Individual
Last updated
Apr 16, 2026
Quick brief
Use AgentShield when an operator needs a preflight security audit of a Claude Code setup, not when they are simply using Claude Code normally. The invoke moment is concrete: scan a .claude directory before enabling a workflow, merging config changes, or trusting a new MCP connection. That scope boundary, static security review of Claude Code configuration, permissions, hooks, and MCP server exposure, keeps this distinct from a generic agent platform or broad security product listing.