The Agent Skills Directory

[SAFE]: The skill instructions explicitly mandate the redaction of secrets and the use of environment variables for credentials, demonstrating a strong security posture for handling sensitive information during the debugging process.
[COMMAND_EXECUTION]: The skill involves the creation and execution of feedback loops such as failing tests, curl scripts, and CLI invocations. These are the primary, documented functions of the skill and are performed locally without requesting elevated privileges or demonstrating malicious intent. The provided bash script template for human-in-the-loop interaction uses standard shell inputs and follows safe coding practices.

diagnosing-bugs