The Agent Skills Directory

[PRIVILEGE_ESCALATION]: The skill instructs the user to execute sudo apt-get install commands to install system-level dependencies like libreoffice and poppler-utils, which involves acquiring administrative permissions.
[REMOTE_CODE_EXECUTION]: The skill requires cloning an external repository (https://github.com/YuhangChen1/Paper2All.git) and installing its dependencies via pip install -r requirements.txt. It then executes several Python scripts from this unverified source (pipeline_all.py, pipeline_light.py, generate_schematic.py), which constitutes downloading and executing remote code.
[COMMAND_EXECUTION]: The skill makes heavy use of the Bash tool to manage the installation process, create Conda environments, and run the transformation pipeline scripts.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process untrusted external data (LaTeX source files and PDFs) which could contain malicious instructions. The skill has significant capabilities including shell execution and file modification, increasing the impact of potential injections.
Ingestion points: The pipeline reads LaTeX directories and PDF files provided as input to the pipeline_all.py and pipeline_light.py scripts.
Boundary markers: No explicit boundary markers or instructions to ignore embedded commands within the paper content are present.
Capability inventory: The skill uses Bash (command execution), Write (file creation), and Edit (file modification) tools.
Sanitization: The documentation does not describe any sanitization, validation, or filtering of the content extracted from the academic papers before it is processed by the LLM or executed in the pipeline.

paper-2-web