The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent on how to use tmuxinator, a utility that executes arbitrary shell commands defined in configuration files through lifecycle hooks (such as on_project_start, on_project_stop, and pre_window) and pane command lists. Example commands in the documentation include sudo systemctl start nginx, docker-compose, and package manager commands.\n- [PROMPT_INJECTION]: The skill describes support for 'Dynamic Configuration (ERB)', which allows arbitrary Ruby code execution during the configuration processing phase (e.g., <%= ENV["USER"] %>). This presents a dynamic code execution surface if the agent handles templates from untrusted sources.\n- [PROMPT_INJECTION]: The skill defines a surface for indirect prompt injection when an agent is tasked with creating, editing, or debugging configurations provided by third parties.\n
Ingestion points: Loading and processing PROJECT.yml configuration files from various filesystem locations (e.g., ~/.tmuxinator/, .tmuxinator.yml) as described in SKILL.md.\n
Boundary markers: Absent. The instructions do not provide delimiters or logic for the agent to distinguish its own instructions from commands embedded within processed configuration files.\n
Capability inventory: Extensive shell execution capabilities including environment setup, process management, filesystem access, and remote access via SSH patterns.\n
Sanitization: Absent. No input validation or sanitization is specified for the configuration content processed by the tool.

tmuxinator