The Agent Skills Directory

[COMMAND_EXECUTION] (SAFE): The skill utilizes high-privilege tools including run_command, write_file, and edit_file. These tools are necessary for the primary DevOps tasks defined in the skill, such as CI/CD pipeline management and infrastructure provisioning. No malicious use of these commands was found.
[DATA_EXPOSURE] (SAFE): The instructions explicitly warn against the exposure of secrets (e.g., .env files) and mandate the use of secret rotation and runtime injection, demonstrating a security-first posture.
[PRIVILEGE_ESCALATION] (SAFE): The skill includes defensive guidelines such as the 'Root Ban' in Docker containers, ensuring processes run with least privilege.
[INDIRECT_PROMPT_INJECTION] (SAFE): The skill possesses a vulnerability surface as it processes external files and executes commands. However, the instructions focus on structured DevOps workflows and safety checks, which mitigates risk within its intended scope.
Ingestion points: Project source code and configuration files via read_file and edit_file.
Boundary markers: None explicitly defined in the provided markdown templates.
Capability inventory: run_command, write_file, edit_file, and read_file.
Sanitization: Not explicitly mentioned, though the skill encourages automated verification pipelines.

resonance-devops