The Agent Skills Directory

[DYNAMIC_EXECUTION]: The skill generates and executes JavaScript files (e.g., build_<name>.js) and Python scripts to automate the creation of PowerPoint presentations. The workflow involves writing data to local JSON files and then reading that data into generated scripts for chart rendering.- [COMMAND_EXECUTION]: The workflow requires the agent to execute shell commands, specifically node and python, to run dynamically created scripts and verification tools like check.py and render.py.- [INDIRECT_PROMPT_INJECTION]: The skill processes data from external sources such as web searches and SEC filings, which creates a surface for indirect prompt injection.
Ingestion points: External data is ingested through WebSearch, WebFetch, and get_sec_filing tools.
Boundary markers: The instructions do not specify any boundary markers or delimiters to separate untrusted external content from the agent's internal instructions.
Capability inventory: The skill utilizes execute_code for Python processing and shell execution for Node.js (node) and Python-based verification scripts.
Sanitization: There are no mentioned mechanisms for sanitizing or validating the content retrieved from external web sources before it is used in the presentation generation process.- [EXTERNAL_DOWNLOADS]: The skill uses WebSearch and WebFetch to retrieve information from general internet sources to populate segments of the company profile, such as brand colors and recent news.

company-profile