The Agent Skills Directory

[COMMAND_EXECUTION]: The skill performs dynamic command execution by invoking the pdflatex compiler on LaTeX files created at runtime. This behavior is documented in the workflow steps in SKILL.md and QUICKSTART.md (e.g., pdflatex -interaction=nonstopmode resume.tex).
[PROMPT_INJECTION]: The skill presents an indirect prompt injection surface because it processes untrusted user data and interpolates it into a LaTeX template that is subsequently executed by a compiler. Ingestion points: User-provided personal, educational, and professional data collected via the agent. Boundary markers: Absent; user input is embedded directly into the .tex source file without isolation or special delimiters. Capability inventory: The skill has the capability to execute system commands via the LaTeX compilation process. Sanitization: Although QUICKSTART.md includes instructions for escaping standard LaTeX special characters to prevent compilation failures, it lacks a robust security sanitization layer to prevent the injection of malicious LaTeX macros or system-level exploitation via the compiler.

resume-builder