The Agent Skills Directory

[SAFE]: The skill is a structured debugging methodology that promotes security best practices. It explicitly mandates the redaction of secrets before displaying artifacts and encourages keeping credentials within environment variables rather than outputting them.- [SAFE]: The Human-In-The-Loop template script provides a controlled way for the agent to interact with the user for manual reproduction steps, reducing the risk of autonomous unauthorized actions.- [INDIRECT_PROMPT_INJECTION]: The skill possesses an attack surface for indirect prompt injection as it ingests untrusted data from logs, traces, and user inputs to drive debugging actions. However, given the primary purpose of bug diagnosis and the inclusion of redaction instructions, the risk is considered low and inherent to the task.

diagnosing-bugs