The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes legitimate shell commands to perform its core functions. It executes pandoc, xelatex, and pdfinfo for document conversion and validation. It also provides instructions for using lpr for document printing. These operations are essential to the skill's primary purpose and are implemented using safe scripting practices, such as array-based command construction in Bash to prevent word splitting and glob expansion issues.
[DATA_EXPOSURE]: No sensitive file access or hardcoded credentials were found. The skill operates on local markdown files and bibliography databases provided by the user within the project directory.
[EXTERNAL_DOWNLOADS]: The documentation suggests installing standard, well-known tools like Pandoc, MacTeX, and DejaVu fonts via brew. These references target official repositories and package managers, which are considered trusted sources.
[INDIRECT_PROMPT_INJECTION]: The skill processes external Markdown data (.md files) to generate PDFs. While this represents an ingestion surface for untrusted data, the processing is limited to document conversion via Pandoc. No evidence was found of unsafe interpolation of this data into agent instructions or shell commands that could lead to code execution. The risk is mitigated by the use of static LaTeX templates and a specific, audited Lua filter (hide-details-for-pdf.lua).

pandoc-pdf-generation