The Agent Skills Directory

[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection by processing external data.
Ingestion points: Untrusted data is retrieved via firecrawl_scrape and firecrawl_research_read_paper as described in SKILL.md.
Boundary markers: No explicit delimiters or instructions to ignore embedded content are used.
Capability inventory: The agent can perform recursive web searches and scraping based on findings.
Sanitization: There is no mention of filtering or validating external content before processing.
[DATA_EXFILTRATION]: The skill initiates network connections to arbitrary external domains.
The tools firecrawl_search and firecrawl_scrape allow the agent to access any URL, which is a standard functionality for this research-oriented skill but involves contacting non-whitelisted domains.

firecrawl-research-index