The Agent Skills Directory

[SAFE]: The skill is designed for read-only diagnostics, focusing on data attribution and reporting. It explicitly prohibits making corrections to data, running operational tasks, or suppressing alerts, which ensures the integrity of the analyzed systems.
[SAFE]: The instructions include robust credential handling guidelines, directing the agent to use environment variables or interactive prompts rather than passing sensitive keys as command arguments or in chat.
[SAFE]: Command execution is limited to standard diagnostic tools like the bruin CLI and git. These operations are performed with safety constraints, such as forced row limits on database queries and dry-run options for wide scans.
[SAFE]: External URLs and resource references are limited to the official Bruin vendor documentation and APIs, which are necessary for the skill's functionality and do not introduce third-party risk.

anomaly-investigate