The Agent Skills Directory

[COMMAND_EXECUTION]: The skill contains multiple scripts that execute system commands to perform automation tasks.
scripts/validate_playbook.py uses the subprocess module to run ansible-playbook --syntax-check and ansible-lint for playbook verification.
scripts/test_playbook.sh utilizes docker CLI commands to manage test containers, including creating, starting, and executing commands within them.
[EXTERNAL_DOWNLOADS]: The testing utility scripts/test_playbook.sh performs network operations to set up the test environment.
It pulls the ubuntu:22.04 image from Docker Hub.
It installs system packages (python3, pip, openssh-client) via apt-get and the ansible package via pip inside the container.
[INDIRECT_PROMPT_INJECTION]: The skill instructions prompt the agent to read existing project files to understand the environment, creating a potential surface for indirect injection if project files are malicious.
Ingestion points: Reads ansible.cfg, inventory/, roles/, and group_vars/ from the project directory (SKILL.md).
Boundary markers: None explicitly defined in the instructions to separate project data from agent instructions.
Capability inventory: File system access (Read/Write), subprocess execution (ansible-playbook), and container management (docker).
Sanitization: The skill includes a dedicated security scanner (scripts/secure_playbook.py) to check for hardcoded credentials and unsafe module usage in generated content.

creating-ansible-playbooks