paper-compile
Warn
Audited by Gen Agent Trust Hub on Nov 10, 2026
Risk Level: MEDIUMPRIVILEGE_ESCALATIONCOMMAND_EXECUTIONDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [PRIVILEGE_ESCALATION]: The skill documentation includes setup instructions that use
sudo apt-get install. An autonomous agent might attempt to execute these commands directly, which would grant it elevated system permissions. - [COMMAND_EXECUTION]: The skill performs multiple shell operations to manage the build environment, including LaTeX compilation (
latexmk), package management (tlmgr), and system information gathering (which,ls,pdfinfo). It also modifies source files based on the content of compilation logs. - [DYNAMIC_EXECUTION]: Step 6 executes a multi-line Python script via
python3 -cto analyze PDF text and verify page limits. This represents runtime code execution of a script defined within the skill. - [INDIRECT_PROMPT_INJECTION]: The skill exhibits an indirect injection surface as it processes external data to drive its behavior.
- Ingestion points: The agent reads user-provided LaTeX source files (e.g.,
main.tex,references.bib) and thecompile.loggenerated by the compiler. - Boundary markers: There are no explicit delimiters or instructions to treat the content of the LaTeX files or logs as untrusted data.
- Capability inventory: The skill possesses significant capabilities, including executing shell commands and performing automated file edits during the error diagnosis and fixing phase.
- Sanitization: The skill lacks mechanisms to sanitize or validate the content extracted from logs or source files before using that content to inform its file-writing and command-execution logic.
Audit Metadata