audit-expert

This skill provides centralized JSON validation for the AGENT_SUCCESS_CRITERIA environment variable. It prevents injection attacks (CVSS 8.2) through defensive parsing and ensures consistent error handling across all agents. Use it when you need secure, validated parsing of success criteria in your Claude Code agents.

The `audit-expert` skill provides expert-level security auditing and compliance guidance for developers, specializing in code reviews, vulnerability assessments, and regulatory frameworks like SOC 2 and GDPR. It helps conduct security audits, analyze code for vulnerabilities, and ensure infrastructure compliance. Use this skill when you need to review code security, prepare for compliance audits, or assess your application against established security standards.

ClawShield is a security audit skill for OpenClaw installations that scans for vulnerabilities and detects prompt injection patterns. It generates a JSON report and is designed for hardening deployments against common attacks. Developers should use it to proactively assess and monitor the security posture of their local setups.

Arc-shield is an output sanitization skill that scans all agent responses to prevent accidental leaks of secrets, tokens, and PII. It acts as a final filter for sensitive data that an agent might inadvertently include in debugging output, code examples, or file contents. Use this security skill when your agent has access to sensitive resources like vaults or environment variables.