The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill extracts information from external, untrusted PDF documents to automate organization tasks. \n
Ingestion points: The agent is instructed to read content from PDF files using pdftotext or OCR tools (mineru-ocr, legal-ocr) to extract dates (Step 3.5). \n
Boundary markers: The instructions lack explicit directives to use boundary markers or to ignore instructions embedded within the extracted text. \n
Capability inventory: Extracted information is used to rename files via mv and to populate Markdown (trademark-info.md) and YAML (case.yaml) files. \n
Sanitization: The process relies on regex-based matching for specific date formats, but does not provide general sanitization for the extracted content. \n- [COMMAND_EXECUTION]: The skill directs the agent to perform various shell commands for its core functionality. \n
Evidence: Use of cp -p and mv for file management with timestamp preservation, date for system time, and pdftotext along with OCR utilities for data extraction.

new-case