The Agent Skills Directory

[PROMPT_INJECTION]: Indirect prompt injection vulnerability surface. The skill ingests untrusted data from LaTeX source files and compilation logs.
Ingestion points: compile.log, paper/main.tex, paper/references.bib, and paper/sections/*.tex.
Boundary markers: Absent. The agent is instructed to read log files and source files directly without specific delimiters or instructions to ignore embedded agent commands.
Capability inventory: The skill uses Bash(*), Write, Edit, Grep, and Glob tools across all files, enabling it to modify files and execute shell commands based on its diagnosis of the input files.
Sanitization: No explicit sanitization or validation of the LaTeX source or log content is performed before the agent interprets it for error fixing.
[COMMAND_EXECUTION]: Automated execution of shell commands for LaTeX processing (latexmk, pdflatex, bibtex) and document verification (pdfinfo, pdftotext, pdffonts).
[EXTERNAL_DOWNLOADS]: Skill instructions involve installing system packages via apt-get, brew, or conda, and automated commands to install LaTeX packages via tlmgr from official repositories.

aris-paper-compile