The Agent Skills Directory

[COMMAND_EXECUTION]: The skill documents the use of Ansible's shell and command modules, which are designed to execute arbitrary commands on remote hosts as part of automation tasks.
[EXTERNAL_DOWNLOADS]: Reference materials include examples for downloading automation roles and collections via ansible-galaxy, as well as fetching files from remote URLs using the ansible.builtin.get_url module.
[CREDENTIALS_UNSAFE]: The documentation references standard paths for sensitive files, such as SSH private keys (~/.ssh/ansible_ed25519) and encrypted vault files (vars/secrets.yml), providing guidance on their secure management in configuration files.
[PROMPT_INJECTION]: As a tool that processes structured YAML and INI files, the skill describes an interface that could be susceptible to indirect prompt injection if processing untrusted input.
Ingestion points: inventory/hosts.yml, playbook.yml, and requirements.yml files.
Boundary markers: The documentation promotes the use of ansible-lint and --syntax-check to validate file structure and content.
Capability inventory: The skill identifies extensive capabilities including remote command execution, file manipulation, and package management across its various reference files.
Sanitization: Recommends the use of Ansible Vault for encrypting sensitive variables and provides instructions for verifying file integrity using checksums.

ansible