The Agent Skills Directory

[SAFE]: The skill functions as a quality assurance tool designed to monitor and verify agent behavior through assertion-based testing.
[COMMAND_EXECUTION]: The skill includes shell scripts (scripts/agent-qa.sh) that use standard utilities such as cat, grep, and diff, along with the claude CLI. These operations are restricted to the local file system and are standard for development and testing workflows.
[CREDENTIALS_UNSAFE]: The documentation and CI configuration examples correctly utilize environment variables and platform secrets (e.g., ${{ secrets.ANTHROPIC_API_KEY }}) for sensitive data, following security best practices. No hardcoded credentials or secrets were identified.
[DATA_EXFILTRATION]: No unauthorized network requests or data exfiltration attempts were detected. External communication is managed solely through the legitimate claude CLI tool to its intended service provider.
[PROMPT_INJECTION]: There are no instructions intended to override safety filters, bypass system prompts, or extract confidential agent instructions. The skill is specifically designed to enforce and validate compliance with existing agent protocols.

agent-qa-testing