The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill provides instructions to install the @llamaindex/liteparse package from the public npm registry and system-level utilities (LibreOffice, ImageMagick) via package managers like brew or apt.
[COMMAND_EXECUTION]: The skill uses shell commands for environment setup and document processing, including npm, lit, brew, and apt-get.
[DATA_EXFILTRATION]: The skill documents an optional --ocr-server-url flag that allows document content to be sent to a local or remote HTTP server for OCR processing.
[PROMPT_INJECTION]: The skill's primary function is parsing unstructured files (PDF, images, Word, etc.), which creates a surface for indirect prompt injection if the processed files contain malicious instructions.
Ingestion points: The lit parse and lit batch-parse commands ingest local files provided as arguments in SKILL.md.
Boundary markers: There are no instructions for using specific delimiters or system prompts to ignore potential instructions within the parsed document text.
Capability inventory: The skill utilizes shell execution for parsing (lit) and maintenance (npm, apt-get).
Sanitization: The instructions do not include methods for sanitizing or validating the extracted text before it is returned to the agent's context.

liteparse