The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes shell scripts (scripts/ensure_ocr.sh) and system utilities (apt-get, pip, ocrmypdf, pdftotext) to install dependencies and manipulate files. These commands are integral to the skill's primary function of PDF processing.
[EXTERNAL_DOWNLOADS]: The installation script fetches software from standard package registries. It installs the ocrmypdf Python package via pip and several system dependencies (tesseract-ocr, ghostscript, pngquant, poppler-utils) via apt-get to enable OCR capabilities. These are well-known tools from trusted registries.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to process untrusted external data in the form of PDF files, which is an inherent risk for AI agents.
Ingestion points: The skill reads input PDF files (in.pdf) to generate a searchable text layer.
Boundary markers: The instructions do not implement delimiters or provide specific warnings to the agent regarding the handling of potentially malicious instructions embedded within the OCR-extracted text.
Capability inventory: The skill environment allows for software installation, shell command execution, and file system access.
Sanitization: There is no documented sanitization or filtering of the text extracted via OCR before it is returned to the agent's context.

ocring-pdfs