The Agent Skills Directory

[SAFE]: The skill utilizes a local Bash script (scripts/init.sh) to automate the creation of project directories and placeholder files. The script follows best practices by using set -euo pipefail and proper variable quoting to prevent command injection or path traversal via user-supplied arguments like project names.
[SAFE]: No network operations or external downloads occur at runtime. While the documentation provides URLs to official style files from reputable academic organizations (e.g., CVPR, ICLR, ACL), the skill correctly instructs the user to download and add these files manually, avoiding automated remote code execution risks.
[SAFE]: The provided CLAUDE.md file serves as a stylistic and technical guide for AI agents working on the project. It establishes constraints for writing style (e.g., active voice, present tense) and ensures the use of predefined LaTeX macros. These constraints are task-specific and do not interfere with the agent's core safety or operational protocols.
[SAFE]: The skill does not access sensitive system directories, hardcode any credentials, or attempt to establish persistence on the host system. The use of Git is limited to local repository initialization within the project folder.

init-latex-project