The Agent Skills Directory

[SAFE]: The skill establishes a safe, report-only framework for technical root-cause analysis. It contains explicit constraints preventing the agent from performing code edits, configuration changes, or environment mutations. Furthermore, it instructs the agent to avoid including secrets in its reports by referencing their locations instead. While the skill is designed to process untrusted data sources like logs and traces, the risk of indirect prompt injection is effectively mitigated by the total absence of system-modifying or code-execution capabilities.

ss-inspect