The Agent Skills Directory

[PROMPT_INJECTION]: The skill processes untrusted content from the internet through search and scraping tools, creating a surface for indirect prompt injection.
Ingestion points: Web data is ingested via the tools firecrawl_search, firecrawl_scrape, web_search_exa, and crawling_exa as described in SKILL.md.
Boundary markers: The workflow instructions do not define delimiters or specific warnings to the agent to treat the scraped content as data only and ignore instructions embedded within it.
Capability inventory: The agent is instructed to launch parallel research subagents and output synthesized reports to the chat or files, which could be influenced by malicious content.
Sanitization: The skill lacks validation or sanitization mechanisms for the data retrieved from external search and crawling results.

deep-research