The Agent Skills Directory

[SAFE]: The skill is designed with a strong emphasis on evidence-based verification and user control. It explicitly prohibits the agent from faking tool outputs and requires manual diff reviews for any automated fixes.
[COMMAND_EXECUTION]: The skill utilizes ansible-lint and ansible-lint --fix to perform its primary function. These commands are executed locally and their usage is clearly documented in the workflow and execution contract.
[DATA_EXPOSURE_AND_EXFILTRATION]: The skill contains explicit instructions not to collect, upload, or store user credentials or code, and it limits operations to the local repository.
[INDIRECT_PROMPT_INJECTION]: While the skill ingests Ansible configuration files and playbooks, the risk is mitigated by the instruction to use standard linters and requiring user confirmation for any changes made to the codebase.

codeguard-ansible