skills/agentskillexchange/skills/Scan Claude Code configs for secrets permission drift and unsafe MCP hookups with AgentShield/Socket
Scan Claude Code configs for secrets permission drift and unsafe MCP hookups with AgentShield
Pass
Audited by Socket on Oct 4, 2026
Checks
Malicious behaviorInjection, exfiltration, untrusted installs
Security concernsCredential exposure, tool/trust exploitation
Code obfuscationHidden or obfuscated code
Suspicious patternsReconnaissance, excessive autonomy, resource use
Audit Metadata