The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to execute shell commands (Python and PowerShell) to perform data analysis tasks and generate Word documents. These commands involve executing local scripts with parameters such as ASIN and dates derived from the agent context.\n- [PROMPT_INJECTION]: Indirect Prompt Injection Surface: The compare_versions.py script reads keyword data from external Excel files via the load_pd_sheet function. This data is processed and formatted into a stats summary which the agent then uses to generate 'insights' and strategy recommendations. Maliciously crafted text within the Excel cells could influence the agent's behavior during report generation.\n
Ingestion points: load_pd_sheet in scripts/compare_versions.py reads data from .xlsx files.\n
Boundary markers: No explicit boundary markers or instructions to disregard embedded commands are provided to the agent when processing the keyword summary.\n
Capability inventory: Shell command execution and file system write access are available to the skill.\n
Sanitization: No validation or sanitization of keyword data is performed before it is summarized for the agent.\n- [EXTERNAL_DOWNLOADS]: The skill requires several standard third-party libraries: pandas, openpyxl, and python-docx. These are well-known packages for data analysis and document creation sourced from official registries.

sif-keyword-tracker