The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and review external Ansible project files and user-defined tasks, which creates a potential surface for indirect prompt injection from malicious project content.
Ingestion points: Processes project-specific files such as ansible.cfg, inventory.yml, and role definitions (SKILL.md).
Boundary markers: No specific boundary markers or "ignore" instructions are defined within the skill to isolate processed data from the agent's logic.
Capability inventory: The skill is authorized to run ansible-lint and ansible-playbook --syntax-check (SKILL.md).
Sanitization: While the skill itself does not sanitize input, it contains explicit instructions for the agent to recommend sanitization (e.g., using | quote and argv for shell commands) in the code it reviews (references/security.md).
[COMMAND_EXECUTION]: The skill directs the agent to execute shell-based diagnostic tools (ansible-lint and ansible-playbook --syntax-check) to validate changes during the review process. This is an expected and benign behavior for a development-focused skill.
[EXTERNAL_DOWNLOADS]: The instructions recommend the use of ansible-galaxy install to fetch project dependencies defined in a requirements.yml file (references/inventory-layout.md). This is a standard workflow for managing Ansible collections and roles from external registries.

pflow-ansible