The Agent Skills Directory

[SAFE]: The skill is designed for technical project maintenance within the RudderStack Profiles framework. Its instructions are consistent with legitimate administrative and development workflows.\n- [COMMAND_EXECUTION]: The skill employs the 'pb' CLI tool for project-specific operations such as 'pb compile', 'pb show', 'pb run', and 'pb migrate'. These are essential domain-specific commands for project validation and materialization and do not pose a security risk in this context.\n- [EXTERNAL_DOWNLOADS]: The skill documentation refers to 'profiles_mlcorelib>=0.8.1' as a necessary Python dependency for ML-related models. This is a vendor-owned package (rudderlabs) and its use is a standard requirement for the features described.\n- [DATA_EXFILTRATION]: No patterns of unauthorized network requests or access to sensitive local system files (e.g., credentials or SSH keys) were identified. All data operations are localized to the profiles project and the configured data warehouse.\n- [PROMPT_INJECTION]: The skill does not contain instructions that attempt to override safety guardrails or manipulate agent behavior. It provides clear, procedural guidance for updating project files.

rudder-profiles-update