The Agent Skills Directory

[COMMAND_EXECUTION]: The render.py script invokes LaTeX compilation tools (latexmk, xelatex) using the subprocess.run method. It correctly passes arguments as a list, which prevents shell injection vulnerabilities by avoiding the use of the shell for command interpretation.
[PROMPT_INJECTION]: The skill processes untrusted input from resumes and job descriptions. It mitigates the risk of indirect prompt injection or document structure corruption by utilizing a robust escape_latex function that sanitizes special characters like backslashes, braces, and percent signs before they are included in the generated .tex file.
[EXTERNAL_DOWNLOADS]: The count_tokens.py script includes an optional dependency on the official anthropic Python package. This package is used exclusively for accurate token usage estimation and is handled according to standard development practices for interacting with well-known services.

cover-letter