The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to gather market and technical evidence from external websites, which could contain malicious instructions designed to influence the agent's behavior.
Ingestion points: The skill uses tavily_search and context_dev_scrape_markdown via the sandbase_call_tool interface to read external pages as defined in references/sandbase-api-map.md.
Boundary markers: No explicit delimiters or instructions are provided to help the agent distinguish between its core instructions and potentially adversarial content within the scraped data.
Capability inventory: The skill relies on network-enabled tools to perform web retrieval and scraping.
Sanitization: The skill does not define any validation or filtering requirements for data returned by the search and scraping tools.
[EXTERNAL_DOWNLOADS]: The skill references and utilizes external tools to fetch market evidence and documentation.
It employs tavily_search for market research and context_dev_scrape_markdown for reading technical pages, which are standard utilities for information gathering within this use case.

prd