The Agent Skills Directory

[SAFE]: The skill is strictly analytical, focusing on the evaluation of failover mechanisms through static design documentation. It includes explicit 'Core Constraints' and 'Execution Rules' that forbid the agent from accessing production systems, reading credentials, or performing any real-world actions like fault injection.
[SAFE]: No evidence of prompt injection or adversarial instructions was found. The skill instructions reinforce professional safety standards by requiring the agent to distinguish between confirmed facts and human-required decisions.
[SAFE]: The skill does not incorporate any external dependencies, remote scripts, or executable code, significantly reducing its attack surface. All provided files are configuration, evaluation cases, or markdown instructions.
[SAFE]: No obfuscation or data exfiltration patterns were detected. All instructions and examples use clear, plain language without hidden encoding or suspicious network operations.

failover-testing