The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructs the agent to download scientific data files from well-known academic and research platforms, including Materials Cloud (materialscloud.org), PseudoDojo (pseudo-dojo.org), and Quantum Espresso (quantum-espresso.org).
[CREDENTIALS_UNSAFE]: The skill correctly handles sensitive information by advising the agent to retrieve API keys from environment variables (e.g., MP_API_KEY) rather than hardcoding them within scripts or instructions.
[PROMPT_INJECTION]: The skill processes data from external sources, creating a surface for potential indirect prompt injection.
Ingestion points: External scientific files (CIF, UPF, PDF, ZIP) and web pages retrieved via WebFetch and Playwright from academic and research sources.
Boundary markers: The instructions do not specify explicit delimiters or warnings for the agent when processing the content of downloaded files to prevent interpretation of embedded instructions.
Capability inventory: The agent is authorized to use Bash, Write, WebSearch, and WebFetch tools, allowing for file manipulation and network access.
Sanitization: The skill assumes the integrity of data from authoritative scientific sources and does not define specific sanitization routines for file contents prior to parsing.

resource-acquisition